Growth and optical properties of
isotopically enriched ZnO nanorods
Ciarán Gray, BSc (Hons)
Thesis Submitted for the Award of
Doctor of Philosophy
School of Physical Sciences
Dublin City University
Research Supervisor
Prof. Enda McGlynn
January 2016
ii
Declaration
I hereby certify that this material, which I now submit for assessment on the
programme of study leading to the award of Doctor of Philosophy, is entirely my
own work, and that I have exercised reasonable care to ensure that the work is
original, and does not to the best of my knowledge breach any law of copyright, and
has not been taken from the work of others save and to the extent that such work has
been cited and acknowledged within the text of my work.
Signed: ____________________
ID No.: 57435622
Date: ______________________
iii
Acknowledgements
This thesis is the culmination of the work of several years, and I wish to
record my sincere thanks to those who helped me along the way, and contributed to
its completion.
First and foremost, thanks to my research supervisor, Prof. Enda McGlynn, to
whom I am indebted for taking me on as a PhD student and giving me the
opportunity to carry out this research. His expertise and insight into the world of
ZnO has been invaluable, as well as his constant encouragement and enthusiasm for
my research.
Thanks to all members of the DCU School of Physical Sciences and the
NCPST who have helped me during this work. In particular I wish to thank Prof.
Martin Henry for his input and sharing of his knowledge of ZnO and Prof. Greg
Hughes for his input on the SIMS. Thank you to Mr. Conor Byrne and Dr. Rajani
Vijayaraghavan for their help with SIMS and Raman Measurements. I would also
like to thank Dr. Daragh Byrne for always answering my questions and passing on
his ZnO expertise.
I am also very grateful for the assistance of all our external collaborators, in
particular Mr. Lukas Trefflich and Prof. Carsten Ronning (Jena, Germany), and Dr.
Joseph Cullen (Linköping, Sweden).
Thanks also to all the technical and administrative staff in the School,
particularly Pat Wogan, Alan Hughes, Des Lavelle and Lisa Peyton.
Thanks to everyone in my lab and office for their advice, help and friendship
over the last few years. They are too many to name all, but I will mention Ruth,
Séamus, Joe, Jack, Damien and Cathal. In addition, I would like to thank Mr.
Saikumar Inguva, for many conversations on ZnO, and for his assistance and
encouragement.
I am eternally grateful to my parents John and Marina, my sister Cróna,
brother Eóin and my fiancée Emer for their constant love and support throughout my
iv
journey. It would not have been possible without them. My thanks are also due to all
my friends, in particular Adam and Jen, for their support and friendship.
Finally, my appreciation is also due to the Irish Research Council for funding
this work through an EMBARK Postgraduate Scholarship. Needless to say, this
work would not have happened without this financial support.
v
Table of Contents
Declaration ............................................................................................................... ii
Acknowledgements ................................................................................................ iii
Table of Contents ..................................................................................................... v
List of Acronyms and Symbols ............................................................................... ix
List of Figures ......................................................................................................... xi
List of Tables .......................................................................................................... xv
Publications .......................................................................................................... xvii
Conferences ........................................................................................................ xviii
Other Presentations and Awards ........................................................................... xix
Abstract ................................................................................................................. xxi
Chapter 1: Introduction to ZnO................................................................................ 1
1.1 Introduction and Applications ....................................................................... 1
1.2 Structural Properties ...................................................................................... 3
1.3 Electronic Properties ..................................................................................... 5
1.4 Growth Techniques ....................................................................................... 8
1.5 Defects in ZnO and the Isotope Effect ........................................................ 12
1.6 Motivation and Objectives .......................................................................... 13
1.7 Outline of Remainder of Thesis .................................................................. 16
1.8 References ................................................................................................... 17
Chapter 2: Growth and Characterisation Techniques ............................................ 24
2.1 Introduction ................................................................................................. 24
2.2 Growth of Zn Isotope-enriched Nanorods ................................................... 25
2.2.1 Substrate Preparation ........................................................................... 25
2.2.2 Seed Layer Preparation ........................................................................ 26
2.2.3 Chemical Bath Deposition ................................................................... 26
2.2.4 Vapour Phase Transport ....................................................................... 30
2.3 Growth of O Isotope-enriched ZnO Nanorods ............................................ 32
2.3.1 Method 1: Modified VS CTR-VPT method in DCU ........................... 33
2.3.2 Method 2: VLS VPT method in Jena ................................................... 34
2.4 Characterisation Techniques ....................................................................... 36
2.4.1 Scanning Electron Microscopy ............................................................ 36
vi
2.4.2 X-ray Diffraction .................................................................................. 40
2.4.3 Secondary Ion Mass Spectroscopy ...................................................... 42
2.4.4 Raman Spectroscopy ............................................................................ 45
2.4.5 Photoluminescence ............................................................................... 47
2.4.6 Reflectance Spectroscopy .................................................................... 56
2.5 References ................................................................................................... 57
Chapter 3: Background Theory of Elements of Optical Properties of ZnO .......... 60
3.1 Introduction ................................................................................................. 60
3.2 Low Temperature Photoluminescence ........................................................ 61
3.3 Defects in Semiconductors .......................................................................... 64
3.4 Structure of the ZnO PL emission spectrum ............................................... 67
3.5 The Cu-related Defect at 2.86 eV ................................................................ 69
3.6 The Isotope Effect ....................................................................................... 74
3.7 Phonon modes ............................................................................................. 76
3.8 Conclusions ................................................................................................. 78
3.9 References ................................................................................................... 78
Chapter 4: Zn Isotope-enriched ZnO Nanorods..................................................... 82
4.1 Introduction ................................................................................................. 82
4.2 Morphology - SEM ..................................................................................... 85
4.3 Alignment and Crystal Quality - XRD ........................................................ 90
4.4 Isotopic Enrichment - SIMS ........................................................................ 94
4.5 Phonon Frequencies - Raman ...................................................................... 98
4.6 Optical Quality and Enrichment - Low temperature PL ........................... 101
4.7 FX Energies - Reflectance Spectroscopy .................................................. 106
4.8 Conclusions ............................................................................................... 107
4.9 References ................................................................................................. 108
Chapter 5: O Isotope-enriched ZnO Nanorods .................................................... 111
5.1 Introduction ............................................................................................... 111
5.1.1 Initial Experiments regarding O-enriched Material ........................... 112
5.1.2 Method 1: Modified VS CTR-VPT method ...................................... 114
5.1.3 Method 2: VLS VPT method ............................................................. 115
5.2 Morphology – SEM ................................................................................... 116
5.3 Alignment and Crystal Quality – XRD ..................................................... 119
vii
5.3.1 Note on Strain .................................................................................... 122
5.4 Isotopic Enrichment – miniSIMS .............................................................. 126
5.5 Phonon Frequencies and Enrichment – Raman ......................................... 128
5.6 Optical Quality and Enrichment - Low temperature PL ........................... 131
5.7 Conclusions ............................................................................................... 134
5.8 References ................................................................................................. 135
Chapter 6: Optical Properties and the Cu-related Defect in Isotopically Enriched
ZnO 137
6.1 Introduction ............................................................................................... 137
6.2 Zn-enriched Nanorods ............................................................................... 138
6.2.1 PL Spectra .......................................................................................... 138
6.2.2 Discussion on Energy Shifts and Involvement of Zni or VO in the Cu-
related Emission at 2.86 eV ............................................................................. 140
6.3 O-enriched Nanorods ................................................................................ 145
6.3.1 PL Spectra .......................................................................................... 145
6.3.2 Energy Shifts ...................................................................................... 147
6.3.3 Discussion on Energy Shifts, Line Widths and Involvement of Oi or
VZn in the Cu-related Emission at 2.86 eV ....................................................... 151
6.3.4 Estimate of Level of Enrichment in O-enriched ZnO ........................ 153
6.4 Discussion on Possible Causes of Reduced Oxygen Enrichment ............. 155
6.5 Conclusions ............................................................................................... 158
6.6 References ................................................................................................. 159
Chapter 7: Conclusions and Future Work ............................................................ 161
7.1 Conclusions ............................................................................................... 161
7.2 Future Work .............................................................................................. 164
7.3 References ................................................................................................. 166
Appendix A: Some Applications of these Growth Methods and Materials in other
Experiments............................................................................................................... A1
A.1 Introduction ................................................................................................... A1
A.2 TRPL of Zn-enriched ZnO nanorods ............................................................ A2
A.3 Third harmonic UV generation with ZnO nanorods ..................................... A4
A.4 Particle and X-ray generation using ZnO nanorods ...................................... A6
A.5 References ..................................................................................................... A7
viii
Appendix B: Efficiency of the SPEX grating and PMT .......................................... B1
B.1 References ...................................................................................................... B1
Appendix C: XRD Strain measurements for the Zn-enriched and VLS O-enriched
samples. ..................................................................................................................... C1
ix
List of Acronyms and Symbols
A0X Acceptor bound exciton
Al Aluminium
Al2O3 Sapphire
Ar Argon
Au Gold
BX Bound exciton
C Carbon
CCD Charge-coupled device
CO Carbon monoxide
CO2 Carbon dioxide
CBD Chemical bath deposition
CCD Charge-coupled device
CTR-VPT Carbothermal reduction vapour phase transport
Cu Copper
CVD Chemical vapour deposition
D0X Donor bound exciton
D+X Ionised donor bound exciton
DAP Donor acceptor pair
DCU Dublin City University
(DI)-H2O (Deionised) water
EDX Energy-dispersive x-ray spectroscopy
FT Fourier transform
FWHM Full width half maximum
FX Free exciton
Ga Gallium
GaN Gallium nitride
H2SO4 Sulphuric acid
H Hydrogen
He Helium
HeCd Helium cadmium
Hg Mercury
x
HMT Hexamethylenetetramine
In Indium
LA Longitudinal acoustic
LED Light emitting diode
LO Longitudinal optical
MBE Molecular beam epitaxy
MFC Mass flow controller
MOCVD Metal oxide chemical vapour deposition
N Nitrogen
NaOH Sodium hydroxide
O Oxygen
Oi Oxygen interstitial
PL Photoluminescence
PLD Pulsed laser deposition
PMT Photomultiplier tube
PTFE Polytetrafluoroethylene
sccm Standard cubic centimetres per minute
SEM Scanning electron microscopy
Si Silicon
SIMS Secondary ion mass spectroscopy
SiO2 Silicon dioxide
SGB Structured green band
SX Surface exciton
TA Transverse acoustic
TCO Transparent conducting oxide
TES Two electron satellite
Ti Titanium
TO Transverse optical
TRPL Time resolved photoluminescence
UV Ultraviolet
VZn Zinc vacancy
VO Oxygen vacancy
VLS Vapour-liquid-solid
VS Vapour-solid
xi
VPT Vapour phase transport
XRD X-ray diffraction
Zni Zinc interstitial
Zn Zinc
ZnO Zinc oxide
ZPL Zero phonon line
List of Figures
Figure 1.1: (a) Hexagonal wurtzite crystal structure of ZnO and (b) the wurtzite unit
cell (yellow cirlces represent oxygen).......................................................................... 3
Figure 1.2: The main crystal planes in wurtzite ZnO.................................................. 4
Figure 1.3: Schematic of the ZnO electronic band structure. ..................................... 6
Figure 1.4: Examples of ZnO nanostructure morphologies from the literature: (a,b)
nanorods60,61
, (c) nanowalls62
, (d) nanodisks63
, (e) nanobelts/nanohelixes64
and (f)
nanobowls.65
................................................................................................................. 9
Figure 2.1: (a) Schematic diagram, and (b) photograph, of the experimental setup
used for the CBD growth step. ................................................................................... 27
Figure 2.2: (a) Schematic diagram of the furnace setup used for the VPT growth of
ZnO nanorods, and (b) photograph of the VPT furnace. ........................................... 31
Figure 2.3: Schematic diagram of the experimental setup for the growth of O-
enriched ZnO nanorods using a modified VS technique. .......................................... 33
Figure 2.4: Schematic diagram showing the general experimental setup for the
growth of O-enriched ZnO nanorods using the VLS technique in Jena. ................... 35
Figure 2.5: (a) Schematic diagram showing the main components of a typical SEM
system and (b) a more detailed diagram showing the secondary electron (SE) and
backscattered electron (BSE) detectors and the associated electronics. (c) Diagram of
the process of generation of EDX signal, (d) Photograph of the Carl-Zeiss SEM
system in DCU. .......................................................................................................... 39
Figure 2.6: (a) Schematic diagram of the layout of an XRD system with zoom
showing diffraction from adjacent crystal planes and the geometry of the Bragg
xii
equation, and (b) photograph of the Bruker AXS D8 Advance Texture
Diffractometer. ........................................................................................................... 41
Figure 2.7: (a) Schematic diagram showing the sputtering process used in SIMS, (b)
photograph of the SIMS system and (c) photograph of the Millbrook miniSIMS
alpha system. .............................................................................................................. 44
Figure 2.8: Energy level diagram showing Rayleigh scattering, and Raman Stokes
and anti-Stokes scattering processes. ......................................................................... 46
Figure 2.9: (a) Schematic, and (b) photograph, of the optical setup used for PL with
the SPEX monochromator.......................................................................................... 48
Figure 2.10: (a) Schematic, and (b) photograph, of the optical setup used for PL with
the FT spectrometer.................................................................................................... 51
Figure 2.11: (a) Schematic, and (b) photograph, of the optical setup used during
reflectance spectroscopy experiments with the FT spectrometer. ............................. 56
Figure 3.1: Diagram showing the excitation and recombination processes in (a)
direct, and (b) indirect, band gap semiconductors. .................................................... 62
Figure 3.2: (a) Schematic of a FX and D0X; (b) Representation of the main
transitions with PL: i) FX, ii) BX, iii) electron to acceptor, iv) donor to acceptor, v)
hole to donor. ............................................................................................................. 66
Figure 3.3: Diagram showing the main types of bound exciton in the band edge
region of ZnO ............................................................................................................. 67
Figure 3.4: Configurational coordinate diagram describing the ground and excited
states of an impurity and the absorption and emission spectra due electron-phonon
coupling. ..................................................................................................................... 72
Figure 3.5: Illustration of how the phonon sideband shape evolves with increasing S
parameter .................................................................................................................... 73
Figure 3.6: E2low
, E2high
, E1 and A1 phonon modes in wurtzite ZnO. ......................... 78
Figure 4.1: Graphical representation of the Zn isotopic content of the eight
Zn-enriched samples. ................................................................................................. 82
Figure 4.2: Typical buffer layer of CBD deposited nanorods. ................................... 83
Figure 4.3: VPT-grown ZnO nanorods with (a) 60 mg and (b) 10 mg of each source
powder at 900°C. ........................................................................................................ 85
Figure 4.4: SEM images showing typical morphology of the ZnO nanorods: (a) plan
view, (b) 30˚ to the vertical and (c) 90˚ to the vertical. Images are from the 64
ZnO
sample. ....................................................................................................................... 86
xiii
Figure 4.5: (a) 90˚ angle of 66/68
ZnO nanorods with CBD buffer layer visible; (b)
64/66ZnO nanorods with CBD buffer layer visible at 90˚; (c) CBD buffer layer at 90˚
in 64/66
ZnO; and (d) 30˚ view of longer nanorods on part of 66
ZnO which have lost
their alignment and become entangled;...................................................................... 87
Figure 4.6: 64/66/68
ZnO samples at 30˚ (a and c), 90˚ (b and d). The first sample (a
and b) shows poor nanorod morphology. The second sample (c and d) has improved
nanorod morphology. ................................................................................................. 88
Figure 4.7: 2θ-ω spectra of several Zn-enriched ZnO samples dominated by the Si
substrate peak at 69.1˚ and the ZnO peak at 34.4˚. .................................................... 90
Figure 4.8: XRD rocking curves of the 34.4˚ (0002) ZnO peak from the nat
ZnO,
64ZnO,
66ZnO and
68ZnO samples. ............................................................................. 93
Figure 4.9: 2θ-ω diffractograms of CBD buffer layer nanorods and ZnO seed layer
only. (Inset shows rocking curve of ZnO (0002) peak in CBD nanorods.) ............... 94
Figure 4.10: SIMS spectra from each sample in the Zn-enriched set. ....................... 96
Figure 4.11: miniSIMS spectra from each sample in the Zn-enriched set. ................ 97
Figure 4.12: Raman spectra of isotopically enriched 64
ZnO, 66
ZnO and 68
ZnO
nanorods. Inset shows the signal from the Si substrate. .......................................... 100
Figure 4.13: PL spectra of the 66
ZnO sample showing (a) a broad range spectrum
displaying the typical PL emission from ZnO nanorods, (b) the intense UV band
edge emission in detail, and (c) the Cu-related 2.86 eV ZPL and associated SGB
(from annealed portion)............................................................................................ 101
Figure 4.14: Comparison of PL intensities of VPT and CBD nanorods. ................. 103
Figure 4.15: Typical PL spectra of selected enriched ZnO nanorod samples showing
the band edge region including the I9 line. ............................................................... 104
Figure 4.16: Reflectance spectrum of a (a) single crystal ZnO sample showing the
A-exciton and B-exciton clearly, and (b) reflectance spectrum of the 64
ZnO, 66
ZnO
and 68
ZnO samples. A and B label the A- and B-excitons. ..................................... 106
Figure 5.1: Image of CBD buffer following VPT step in oxygen-depleted
atmosphere. No nanorod growth is observed. ......................................................... 115
Figure 5.2: SEM images of sample Zn16
O-VS at (a) ~40° and (b) ~90°, and the
middle of sample Zn16/18
O-VS at (c) ~40° and (d) ~90°. ......................................... 116
Figure 5.3: SEM images of the morphology of the VLS samples with 30° tilt; (a)
Zn16
O-VLS, (b) Zn16
O-VLS showing hexagonal structure, (c) Zn18
O-VLS and (d)
Zn16/18
O-VLS. .......................................................................................................... 118
xiv
Figure 5.4: 2θ-ω diffractograms of the VS O-enriched samples dominated by the Si
substrate peak at 69.1˚ and the ZnO peak at 34.4˚. Inset shows rocking curves of the
ZnO (0002) peak. ..................................................................................................... 119
Figure 5.5: 2θ-ω diffractograms of the VLS O-enriched samples dominated by the
Si substrate peak at 69.1˚ and containing multiple ZnO peaks. Inset shows rocking
curves of the ZnO (0002) peak................................................................................. 122
Figure 5.6: (a) Symmetric and (b) asymmetric XRD reflections, which enables (c)
the mapping along a line in reciprocal space which is not normal to the substrate. 123
Figure 5.7: s-plane and r-plane 2θ peaks (with in-plane component) in the tilted VS
O-enriched samples.. ................................................................................................ 124
Figure 5.8: Typical miniSIMS spectra of the Zn16
O and Zn18
O samples in the VS set
(a and b) and the VLS set (b and d). ........................................................................ 126
Figure 5.9: Raman spectra of the VS method O-enriched ZnO nanorods. Inset shows
the signal from the Si substrate. ............................................................................... 128
Figure 5.10: Raman spectra of the VLS method O-enriched ZnO nanorods. Inset
shows the signal from the Si substrate. .................................................................... 128
Figure 5.11: Effect of the 900°C 10 minute anneal on the green band region. ....... 131
Figure 5.12: Band edge region of the PL spectrum in the (a) VS and (b) VLS O-
enriched samples. ..................................................................................................... 133
Figure 6.1: Band edge (a and b) and Cu-related 2.86 eV ZPL (c and d) spectra from
the Zn-enriched samples from the SPEX system (a and c) and the FT system (b and
d) .............................................................................................................................. 139
Figure 6.2: Energies of the I9 lines (black) and Cu-related ZPL lines (blue) in the Zn-
enriched ZnO samples, from the (a) SPEX and (b) FT systems. ............................. 144
Figure 6.3: Band edge (a and b) and Cu-related 2.86 eV ZPL (c and d) spectra from
the VS and VLS O-enriched samples from the FT system (a, b and d) and the SPEX
system (c). ................................................................................................................ 146
Figure 6.4: Energies of the I9 lines (black) and Cu-related ZPL lines (blue) in the (a)
VS, and (b) VLS, O-enriched ZnO samples, from the SPEX system. ..................... 149
Figure 6.5: Energies of the I9 lines (black) and Cu-related ZPL lines (blue) in the (a)
VS, and (b) VLS, O-enriched ZnO samples, from the FT system. .......................... 150
Figure A.1: TRPL decay curves of (a) the I9 recombination, (b) the I9-1LO line, and
(c) the I9 and SX features, in Zn-enriched ZnO nanorods......................................... A3
Figure A.2: CBD grown ZnO nanorods on a fused silica substrate ......................... A5
xv
Figure A.3: ZnO nanorods on Ti foil grown by the (a) NaOH and (b) HMT chemical
baths. ......................................................................................................................... A6
Figure B.1: SPEX grating and PMT efficiency curves. ........................................... B1
List of Tables
Table 2.1: Natural abundances of Zn isotopes in ZnO and the enrichment levels of
the source powders used to produce Zn-isotopically enriched nanorods. .................. 32
Table 4.1: Set of Zn-enriched ZnO nanorods............................................................ 83
Table 4.2: ZnO (0002) peaks, FWHM and crystallite size and ZnO c lattice
constants, and Si (004) peaks and lattice constants in Zn-enriched nanorods. .......... 91
Table 4.3: Peaks and FWHM of the rocking curves of the 34.4˚ ZnO (0002) peak in
all samples. ................................................................................................................. 93
Table 4.4: Frequencies and FWHM of the E2low
and E2high
phonons for 64
ZnO, 66
ZnO
and 68
ZnO samples. .................................................................................................. 100
Table 4.5: Energies and FWHM of the I9 exciton recombination in samples with
different Zn isotopic enrichments. ........................................................................... 104
Table 5.1: Sets of O-enriched ZnO nanorods successfully produced using both
methods. ................................................................................................................... 112
Table 5.2: ZnO (0002) 2θ peak and FWHM, and calculated crystallite size and c
lattice constant for the VS and VLS O-enriched samples. ....................................... 120
Table 5.3: Peaks and FWHM of the rocking curves of the 34.4˚ ZnO (0002)
reflection the VS O-enriched samples. .................................................................... 120
Table 5.4: s-plane and r-plane 2θ peaks in the VS (with in-plane component) and
VLS O-enriched samples (from standard out-of-plane diffractograms for
comparison). ............................................................................................................. 125
Table 5.5: Frequencies and FWHM of the E2low
and E2high
phonons for the VS and
VLS O-enriched samples. ........................................................................................ 130
Table 5.6: Peak positions and FWHM of the I9 exciton recombination in the VS and
VLS O-enriched samples. ........................................................................................ 133
Table 6.1: Energies and widths of the I9 and ZPL lines in Zn-enriched ZnO, from the
SPEX system. ........................................................................................................... 142
xvi
Table 6.2: Energies and widths of the I9 and ZPL lines in Zn-enriched ZnO, from the
FT system. ................................................................................................................ 142
Table 6.3: Energies and widths of the I9 and ZPL lines in O-enriched ZnO, from the
SPEX system. ........................................................................................................... 147
Table 6.4: Energies and widths of the I9 and ZPL lines in O-enriched ZnO, from the
FT system. ................................................................................................................ 148
Table 6.5: Summary of the band edge and ZPL energy shifts in the O-enriched
samples. .................................................................................................................... 150
Table 6.6: 18
O enrichment as % of literature PL and Raman references for Zn18
O and
Zn16/18
O nanorods. .................................................................................................... 154
Table A.1: Fast (τf) and slow (τs) component time constants of the I9 decay lifetime
and their relative amplitudes in Zn-enriched ZnO nanorods. ................................... A4
Table C.1: s-plane and r-plane 2θ peaks (with in-plane component) in the Zn-
enriched and VLS O-enriched samples. .................................................................... C2
xvii
Publications
Ciarán Gray, Lukas Trefflich, Carsten Ronning, Enda McGlynn. Growth of oxygen-
enriched ZnO nanorods by two novel VPT methods. [In preparation]
Ciarán Gray, Enda McGlynn, et al. Optical properties of isotopically enriched ZnO
nanorods and study of the Cu-related defect at 2.86 eV. [In preparation]
Saikumar Inguva, Ciarán Gray, Enda McGlynn, Jean-Paul Mosnier. Origin of the
3.331 eV emission in ZnO nanorods: comparison of vapour phase transport and
pulsed laser deposition grown nanorods. [Submitted to the Journal of Luminescence]
A. Andreev, K. Platonov , J. Braenzel, A. Lübcke, S. Das, H. Messaoudi, R.
Grunwald, C. Gray, E. McGlynn, M. Schnürer. Relativistic laser nano-plasmonics
for effective fast particle production. Plasma Physics and Controlled Fusion. 58,
014038 (2016).
Ciarán Gray, Joseph Cullen, Conor Byrne, Greg Hughes, Irina Buyanova, Weimin
Chen, Martin O. Henry, Enda McGlynn. Growth of isotopically enriched ZnO
nanorods of excellent optical quality. Journal of Crystal Growth 429 (2015) 6–12.
S.K. Das, F. Güell, C. Gray, D. Byrne, P.K. Das, R. Grunwald, G. Steinmeyer, E.
McGlynn. Comparison of linear and nonlinear optical properties of ZnO nanorods.
Chapter 11 (pages 193-206) of Progress in Nonlinear Nano-Optics (Editors: S.
Sakabe, C. Lienau and R. Grunwald). Springer Verlag, 2015.
xviii
S.K. Das, F. Guell, C. Gray, P.K. Das, R. Grunwald, E. McGlynn.
ZnO nanorods for efficient third harmonic UV generation: erratum.
Optical Materials Express, 4 (2014) 1243.
S.K. Das, F. Guell, C. Gray, P.K. Das, R. Grunwald, E. McGlynn.
ZnO nanorods for efficient third harmonic UV generation.
Optical Materials Express, 4 (2014) 701–709.
Conferences
A. Andreev, K. Platonov , J. Braenzel, A. Lübcke, S. Das, H. Messaoudi, R.
Grunwald, C. Gray, E. McGlynn, M. Schnürer. Relativistic laser nano-plasmonics.
Oral presentation (by A. Andreev) at the 42nd European Physical Society, EPS,
Conference on Plasma Physics, 22-26 June 2015, Lisbon, Portugal.
Enda McGlynn, Frank Güell, Susanta Kumar Das, Ciarán Gray, Daragh Byrne,
Prasanta Kumar Das, Günter Steinmeyer, and Ruediger Grunwald. THG of ZnO
nanorods for efficient third order interferometric FROG. Oral presentation (by Enda
McGlynn) at the Laserlab Europe Users Meeting, 29-30 September 2014, Prague,
Czech Republic.
Julia Braenzel, Matthias Schnürer, Alexander Andreev, Susanta Kumar Das, Hamza
Messaoudi, Ruediger Grunwald, Ciarán Gray, Enda McGlynn. Laser ion acceleration
with optimized foil target morphology and atomic composition. Oral presentation
(by Julia Braenzel) at the European Conference on Laser Interaction with Matter,
ECLIM 2014, 31 August – 5 September 2014, Paris, France.
xix
Susanta Kumar Das, Frank Güell, Ciarán Gray, Daragh Byrne, Prasanta Kumar Das,
Ruediger Grunwald, Enda McGlynn, Günter Steinmeyer. THG of ZnO nanorods for
efficient third order interferometric FROG. Oral presentation (by Günter
Steinmeyer) at the Conference on Lasers and Electro-Optics: Science & Innovations
(CLEO/SI) 2014, 8-13 June 2014, San Jose CA, USA.
C. Gray, M.O. Henry, E. McGlynn. Effects of Zn isotopic enrichment on the Cu-
related 2.86 eV emission in ZnO nanorods. Poster presentation at Photonics Ireland
2013, 4-6 September 2013, Belfast, UK.
Ciarán Gray, Martin O. Henry, Daragh Byrne, Enda McGlynn. Effects of Zn
isotopic enrichment on the Cu-related 2.86 eV emission in ZnO nanorods. Poster
presentation at the International Conference on Defects in Semiconductors, ICDS27,
21-26 July 2013, Bologna, Italy.
C. Gray, D. Byrne, M. Henry, E. McGlynn. Isotope enrichment and substitution in
the study of impurities and defects in ZnO semiconductors. Oral presentation at
Condensed Matter Division of the European Physical Society, CMD-24, 3-7
September 2012, Edinburgh, Scotland, UK.
Other Presentations and Awards
C. Gray, L. Trefflich, J. Cullen, C. Byrne, G. Hughes, I. Buyanova, W. Chen, C.
Ronning, MO. Henry, E. McGlynn. Growth and Optical Properties of Isotopically
Enriched ZnO Nanorods. Poster presentation at NCPST Postgraduate Poster
Competition, Dublin City University, 10 December 2015.
xx
Ciarán Gray, Martin O. Henry, Daragh Byrne, Enda McGlynn. Effects of Zn
isotopic enrichment on the Cu-related 2.86 eV emission in ZnO nanorods. Poster
presentation at NCPST Postgraduate Poster Competition, Dublin City University, 17
December 2013. Runner-up.
C. Gray, D. Byrne, M. Henry, E. McGlynn. Isotope enrichment and substitution in
the study of impurities and defects in ZnO semiconductors. Poster presentation at
BOC School of Physical Sciences Poster Competition, Dublin City University, 31
May 2013.
C. Gray, D. Byrne, M. Henry, E. McGlynn. Isotope enrichment and substitution in
the study of impurities and defects in ZnO semiconductors. Poster presentation at
NCPST Postgraduate Poster Competition, Dublin City Univeristy, 18 December
2012. Winner, Best Early Stage Researcher.
C. Gray, D. Byrne, E. McGlynn, M.O. Henry. Isotope enrichment and substitution
in the study of impurities and defects in ZnO semiconductors. Poster presentation at
BOC School of Physical Sciences Poster Competition, Dublin City University, 13
April 2012.
xxi
Abstract
Growth and optical properties of isotopically enriched ZnO nanorods
Ciarán Gray, BSc (Hons)
The growth and analysis of the optical properties of isotopically enriched ZnO
nanorods is presented. ZnO nanorods were grown on silicon substrates by chemical
bath deposition on a drop-coated ZnO seed layer followed by carbothermal-reduction
vapour phase transport (VPT). Zn-isotopically enriched samples were grown using
very small amounts of Zn-enriched ZnO source powders in the VPT phase. Oxygen-
enriched ZnO nanorods were grown by two modified novel vapour-solid (VS) and
vapour-liquid-solid (VLS) methods. Characterisation by scanning electron
microscopy and x-ray diffraction confirm the growth of high quality, dense, c-axis
aligned nanorods over a large surface area in Zn-enriched and VS O-enriched
nanorods, and non-aligned nanorods for VLS O-enriched samples. Zn isotopic
enrichment was confirmed by secondary ion mass spectroscopy and Raman. Optical
studies by low temperature photoluminescence (PL) show shifts in the band edge
exciton recombination energy with changing isotopic masses, again confirming
enrichment. The optical quality of the nanorods was excellent. The well-known Cu-
related emission at 2.86 eV was studied by PL in order to investigate possible
involvement of intrinsic, native defects such as interstitials and vacancies in this
deep centre. Shifts in this zero phonon line (ZPL) energy were measured and
compared to changes in the band edge energies. No relative shift was observed in
these Zn-enriched samples, indicating only O atoms in the vicinity of the Cu atom,
and no involvement of Zn interstitials or O vacancies in this Cu defect. PL and
Raman shifts in the O-enriched samples were less than those previously reported for
the same nominal enrichment levels, indicating possible lower enrichment than
anticipated. Band edge and ZPL emissions in samples with O-enrichment displayed
a relative shift and the ZPL line widths showed a substantial increase, which could
indicate O interstitials or Zn vacancies complexing with Cu, however we attribute
these effects instead to the multiple local configurations possible for O atoms in
these mixed isotope environments in such lesser enriched samples.
1
Chapter 1: Introduction to ZnO
1.1 Introduction and Applications
This chapter presents an introduction to zinc oxide (ZnO). It begins with a
brief overview of semiconductors in general as well as ZnO, in addition to presenting
a description of the uses and applications of ZnO more recently. Basic material
properties of ZnO, including its structural and electronic properties are discussed.
The main growth techniques used to produce ZnO crystals and nanostructures are
described. Finally, the motivation for the present work and a brief outline of this
thesis is presented.
Since the first semiconductor transistors were developed in the mid-twentieth
century, the industries built on these devices have permeated modern life and been at
the centre of a technological revolution. Almost every modern electronic device that
we use today contains semiconductor transistors. Silicon (Si) has become the
dominant semiconductor material used in the computing and electronics industries.
However Si is not suitable for some other applications, in particular for devices
involving optical emission, because of its indirect band gap. Subsequently, many
other semiconducting materials have received much research attention as their
properties may be more suited to a particular application. ZnO is a semiconductor
which has received very considerable attention from many researchers. This has
2
been driven by the large array of potential applications of the material, both in single
crystal and nanostructure forms, particularly in recent decades.
ZnO is a widely known II-VI wide band gap compound semiconductor
material which has been used and studied for many years. For example, it has
recorded as being used in metallurgy processes over 3000 years ago.1 More recently,
there was much research done on ZnO over the last century with a first peak around
the 1960s, encouraged by improved growth techniques for high quality single crystal
material. The semiconducting properties of ZnO were recognised around the 1920s2
and by mid-century its UV and green emission spectrum were known, as well as
many of its structural and electronic properties.3 ZnO continued to be a popular
material with researchers before experiencing a second peak in output in the
literature in the last twenty years including a large amount related to nanostructure
research as ZnO forms nanostructures with a wide variety of morphologies. This is
unsurprising given the huge interest in nanotechnology for device applications at the
present time. However there are much fewer reports in general on ZnO-based
devices than on nanostructure growth methods and optical properties.
ZnO is a direct band gap material which has received much attention due to its
favourable optical properties for potential use in photonic and optoelectronic
devices. It can also be synthesised using many different methods. It tends to deposit
in nanostructured form and in a wide range of morphologies which can be controlled
using various growth parameters. ZnO has a band gap of 3.3 eV at room
temperature and an exciton binding energy of 60 meV.4 ZnO is therefore a very
efficient emitter in the ultraviolet (UV) region of the spectrum and this exciton
binding energy compares favourably to other materials known for their blue or UV
emissions, for example the figure for gallium nitride (GaN) is 25 meV.5 These
optical properties, which are of special interest in this work, make it attractive for use
in devices such as UV emitters, photodetectors (including photodiodes and
photovoltaics) and UV emitters such as LEDs6 and laser diodes.
7–10 ZnO has also
been used in transparent conducting oxides (TCO) in solar cells and as a transistor
material.11–13
Nanorods in particular allow increased surface area in solar cells
which in turn increases the amount of photons captured and the current produced.
ZnO nanostructures have attracted attention for use in solid state gas sensors due to
their large surface to volume ratio and surface reactivity.14,15
Metal doped ZnO may
3
be ferromagnetic giving rise to potential spintronics applications.16,17
Applications
making use of its piezoelectric properties for energy scavenging and surface acoustic
wave applications have also been reported.18,19
ZnO has been used in cold cathode
applications like field emission in flat panel displays.8,20,21
Ordered ZnO nanorods
arrays in particular have shown much potential in this area.22
ZnO has shown
potential uses in biomedical applications, such as pharmaceuticals, because it is a
biocompatible material.23,24
Additionally, ZnO has been used in cosmetics, sun
cream, food, paints, pigments and rubber materials. Over 100,000 tonnes of ZnO is
produced annually, mostly for use in these lower “tech” applications.3
1.2 Structural Properties
(a) (b)
Figure 1.1: (a) Hexagonal wurtzite crystal structure of ZnO and (b) the wurtzite unit
cell (yellow cirlces represent oxygen). Reproduced from reference 25.
ZnO is usually found in the hexagonal wurtzite form as this is its most stable
phase at standard temperature and pressure and over a wide range of temperatures
and pressures. However ZnO can also crystallize in the cubic zinc-blende or rocksalt
allotropes in some circumstances.26
Grown on cubic substrates, the zinc-blende form
is stable in ambient conditions.27
ZnO can also crystallize in the rocksalt phase at
4
higher pressures around ~9-10 GPa, which is a metastable phase at lower pressures.28
The zinc-blende and rocksalt forms are more common in other II-VI compounds and
oxides. As this work involving ZnO was carried out at or below atmospheric
pressure, wurtzite is the relevant phase here and only this allotrope is discussed in
detail.
The wurtzite form has a hexagonal structure where each zinc (Zn) atom is
surrounded by four oxygen (O) atoms in the shape close to a tetrahedron. Likewise,
each O atom is surrounded by four Zn atoms. This structure is shown in figure
1.1(a). The wurtzite unit cell is described by two lattice parameters, denoted a and c,
as shown in figure 1.1(b). Ideally, i.e. in the case where the coordination is perfectly
tetrahedral, they have an axial ratio of c/a = √(8/3) = 1.6330.3 Experimental
measurements of these lattice parameters in ZnO have been reported as a = 0.32498
nm and c = 0.52066 nm, giving with c/a = 1.6021. However there are reports of this
ratio varying over a small range. For example, the figure has been given as 1.602129
and 1.59030
, and also varying over a similar range with pressure.31
The observed
variations from the ideal structure may be due to the partially ionic nature of the ZnO
bond which could lead to a distortion in the bonding angle.7 The volume of the
wurtzite ZnO unit cell is 23.8 x 10-3
nm3.32
Figure 1.2: The main crystal planes in wurtzite ZnO. Reproduced from reference
33.
5
The main crystal planes of wurtzite ZnO are shown in figure 1.2.33
The
structure along the c-axis consists of alternating close packed planes containing Zn
or O atoms, with each atom coordinated with four atoms of the other type in the
close packed tetrahedron structure described above. This pattern leads to polar
symmetry along the c-axis as the c-plane is terminated with either Zn or O atoms.
The c-plane is denoted (0001) for the positive planes (Zn-terminated) and (000-1) for
the negative planes (O-terminated) using the usual Miller-Bravais notation. The a-
plane (1-120) and m-plane (10-10) contain the c-axis and are not polar, as they
contain equal amounts of Zn and O atoms. The semi-polar r-plane (10-11) and R-
plane (10-12) are also shown, although there is less published information in terms
of studies involving these planes and it has proven difficult to achieve epitaxial
growth in these directions.34
The c-plane has the lowest surface energy which gives
rise to preferential nanostructure growth along this direction as new material deposits
more favourable on the polar faces than the non-polar ones.35
This property is
inherent to the growth of ZnO nanorods in this work. This type of crystal structure,
without a centre of inversion, also gives rise to piezoelectric and pyroelectric
properties.
1.3 Electronic Properties
The electronic properties of ZnO are one reason why there has been much
interest in the material. As mentioned above, ZnO is a wide band gap semiconductor
whose band gap at room temperature is 3.3 eV. This band gap energy corresponds to
photon emission in the UV region. The material exciton binding energy of 60 meV4
is larger than the thermal energy kT at room temperature. This is an important
parameter as is means that many correlated electron-hole pairs, called excitons, can
exist up to, and even above, room temperature, and the correlation of the electron
and hole motion leads to closer proximity and a larger optical matrix element. ZnO
is therefore an efficient emitter in the near UV region at room temperature via
exciton recombination, leading to much interest in research in possible
optoelectronic devices based on ZnO.9,36
6
Figure 1.3: Schematic of the ZnO electronic band structure. Reproduced from
reference 37.
Figure 1.3 shows a schematic diagram of the electronic band structure in
ZnO.37
The outer electron configuration in Zn is 3d10
4s2, and the outer
configuration in O is 2s2 2p
4. ZnO displays hybridised sp
3 bonding. The outer 4s
2
electrons in Zn atom enter into partially ionic bonds with the O atom to stabilise the
O outer shell, due to the greater electronegativity of the O atom, and we can view
aspects of the band structure in terms of an ionic bonding perspective. The ZnO 3.3
eV band gap is between the O2-
2p valence band and the Zn2+
4s conduction band
states. The band gap is direct meaning that the conduction band minimum and
valance band maximum are at the same position in reciprocal space (their k-vectors
are the same). The conduction band mainly consists of empty Zn2+
4s orbitals with
Γ7 symmetry while the valence band is made up of the occupied O2-
2p orbitals.
These O p-type orbitals are split by both spin-orbit coupling and crystal field effects
(i.e. a departure from cubic symmetry) into three twofold degenerate (Kramers
degeneracy) bands labelled A, B and C, with symmetries of Γ7, Γ9 and Γ7
respectively, from the highest energy to the lowest, by the tetrahedral crystal field
and spin-orbit coupling.3,37
The ordering of the Γ7, Γ9 and Γ7 bands has been the
subject of great controversy and discussion over the years, but the Γ7, Γ9 and Γ7
ordering (different to other II-VI materials such as cadmium sulfide) seems well
7
established now, and the specific ordering is of little importance for the present
work, so we do not dwell on it. Further detail is given in reference 37, and the
influence of Zn 3d levels are suspected to be the origin of the difference between
ZnO and other II-VI materials. This topic is also related to other discrepancies in
terms of the band structure in ZnO, specifically that some experiments are not totally
explained by theoretical calculations.7 The Zn 3d electrons are treated as core shell
electrons in theoretical models, however experimental results suggest that they may
have an influence on the position of the Zn 4s and O 2p valence band electrons.7,38–41
However, the primary influence on the valence band position is the hybrid orbitals of
the Zn 4s and O 2p electrons as discussed above.
ZnO, in its natural and nominally undoped state, actually in almost all cases,
has a significant abundance of n-type carriers i.e. electrons. A number of
explanations for this property have been proposed in the literature. Deviations in the
material stoichiometry can lead to an excess of Zn interstitials (Zni, an additional Zn
atom where there should be none) or O vacancies (VO, a defect of a missing O atom
surrounded by Zn atoms) in the ZnO lattice. It has been suggested that these native
defects are responsible for this n-type conductivity, but uncertainty remains as to
whether this is due to the interstitials or vacancies.7,42,43
It has also been suggested
that hydrogen (H) is responsible for this conductivity due to the generally ubiquitous
presence of H during the growth of ZnO material, as well as its high
diffusivity.7,42,44,45
Although H is amphoteric in most semiconductors which means
it is in the H+ state in p-type material and in the H
- state in n-type material, it is
always found in the H+ state in ZnO and acts as a donor as would be expected in n-
type ZnO.7,45
The n-type behaviour can also be increased with further intentional
doping with aluminium (Al), gallium (Ga) and indium (In), all of which are present
in trace amounts in the material grown in this work. Further discussion of such
defects in ZnO is contained in chapter 3.
In contrast, p-type doping of ZnO material has proven difficult, due, at least
in part, to its intrinsic n-type state. This type of electronic behaviour is known as
unipolarity. Other potential contributing factors include a low solid solubility of p-
type dopants46
and a large ionisation energies of possible acceptors.2 Indeed many
elements which would be expected to act as acceptors in ZnO do not actually
contribute to p-type conductivity.7 There are some reports in the literature of p-type
8
doping in ZnO using a number of techniques and dopants7,47–53
, but there are no
reports of reliable and stable p-type ZnO being synthesised in a reproducible manner.
This has obvious detrimental consequences for ZnO photonic or optoelectronic
devices which generally rely on a p-n junction to operate, and where stable p-type
material is thus a prerequisite. Indeed attempts to produce p-n junction based
devices of ZnO have proven unsuccessful thus far, at least in terms of reproducible
production at scales beyond once-off laboratory realisations. However, progress has
been made in producing heterojunction devices using n-type ZnO and other p-type
materials, e.g. GaN or organic thin films, among others.6,54–56
1.4 Growth Techniques
ZnO has been produced in a wide variety of morphologies, including single
crystals and a variety of nanostructures, by a multitude of techniques. The most
common ones are discussed briefly in this section. The growth method used will
depend on the specific properties required for a particular application.
Bulk single crystals of ZnO have been synthesised in a number of ways
including vapour phase transport (VPT)43,57
and hydrothermal techniques.58
The
VPT method typically involves the reduction of high quality ZnO to Zn and O2 at
high temperatures in a furnace. The Zn vapour is then transported to a cooler part of
the furnace where it oxidises on the substrate. The temperature is usually in the 850-
1150 °C range, depending on the reducing agent. Although significant sublimation
and decomposition of ZnO requires temperatures greater than 1900 °C, the reducing
agent allows for growth at lower temperatures. The growth time can be several
weeks for larger crystals.59
In hydrothermal growth methods, the transport of Zn
from the source to the substrate is by convection through a solvent solution. Instead
of the reducing agent as in VPT, the method makes use of a mineralisation agent
which dissolves the ZnO into solution. The ZnO is then transported to the cooler
region and precipitates out of the solution forming the crystals.
9
Figure 1.4: Examples of ZnO nanostructure morphologies from the literature: (a,b)
nanorods60,61
, (c) nanowalls62
, (d) nanodisks63
, (e) nanobelts/nanohelixes64
and (f)
nanobowls.65
Nanostructures of ZnO have been of much interest in the literature. ZnO
nanostructures have been produced in the form of nanorods and nanowires12,20,60,61,66
,
nanowalls62
, nanodisks63
, nanohelixes, nanosprings, nanorings, nanobelts64
,
nanobowls65
, nanosheets67
and thin films68
, to name a few. Some examples of these
morphologies are shown in figure 1.4. One popular method often used to deposit
ZnO nanostructures and thin films is chemical bath deposition (CBD).26,60,69
This is
a relatively inexpensive technique that takes place at relatively low temperatures,
usually below 100 °C. This allows for the use a wide variety of substrates which
10
might not be capable of withstanding higher temperatures. Generally a Zn salt such
as Zn acetate, nitrate, chloride or sulphate is used as a precursor and reacted with a
base chemical to form a partially soluble Zn hydroxide.60,70–72
A number of
chemicals have been used as a base including hexamethylenetetramine (HMT),
sodium hydroxide (NaOH), ammonia, ethanolamine and urea.26,69,72,73
The Zn
hydroxide then decomposes into ZnO or Zn2+
which is then oxidised to form ZnO.74
The exact reaction pathway will depend on the particular bath and conditions used.
Solvents used in chemical baths have included water, ethanol, methanol and
propanol.75
The morphology of the final nanostructures or thin films can be
controlled by altering the chemicals used, the solvent, the pH, the temperature,
growth time, substrate and the addition of surfactant. As such, CBD is a flexible and
popular, low temperature method used to deposit ZnO. CBD is used extensively
throughout this work and is described further in chapters 2 and 4.
In addition to bulk single crystals, VPT is also commonly used to deposit
nanostructures.76,77
The source powder can be Zn metal or a ZnO powder mixed
with a reducing agent. As in the bulk crystal case, the Zn vapour produced is carried
by a carrier gas to the substrate where it is oxidised and deposits. The substrate is
usually downstream of the source in the furnace, but can also be beside or above it.
The reducing agent allows for the reduction of ZnO at temperatures of ~900 °C
while direct sublimation of ZnO requires much higher temperatures of greater than
1900°C. Having said this, sufficient evaporation of ZnO occurs at temperatures of
around 1350°C to allow growth.78,79
VPT is used extensively in this work to grow
ZnO nanorods on buffer layers grown by CBD. The reducing agent used here is
graphite. With the addition of carbon (C) the method is known as carbothermal
reduction vapour phase transport (CTR-VPT). There are two main growth pathways
which may be used with VPT. The first is the vapour-solid (VS) method, where Zn
is adsorbed onto the surface and oxidised to form ZnO. In this case the CBD buffer
layers provide energetically favourable nucleation sites on the substrate for
deposition.80
The second method is the vapour-liquid-solid (VLS) method. In this
case a thin layer of a metal catalyst like gold (Au) is deposited on the substrate in
advance of VPT growth.76
The Au is molten at the growth temperature and the Zn
vapour is adsorbed into the Au droplets forming a saturated alloy. The Zn then
precipitates out of the Au droplet and oxidised to form ZnO nanorods. The Au
11
droplets end up at the top of the nanorods after growth. VPT growth is discussed in
more detail in chapters 2, 4 and 5.
There are many alternative ways which have been used to produce ZnO
nanostructures and thin films, including chemical vapour deposition (CVD),
molecular beam epitaxy (MBE) and pulsed laser deposition (PLD).81
CVD involves
exposing the heated substrate to a volatile precursor which is transported to the
substrate by a carrier gas. The precursor decomposes on the substrate and forms the
deposit. When the precursor used is metal organic (MO), this is known as MOCVD.
MOCVD is routinely used for growth of thin films, as well as aligned
nanostructures.82
MBE is widely used in the growth of thin films.2,83
It involves a
number of thermally generated molecular beams interacting on a substrate in a high
vacuum. It has also been used to produce ZnO nanorods.84
PLD has been used to
deposit ZnO films and nanostructures using a pulsed laser beam to vapourise the
surface of a ZnO or Zn metal target in a vacuum.85–87
The resulting plasma expands
in the chamber and is incident on a substrate a small distance away from the target
and the material deposits on the substrate. The properties of the resulting
nanostructure can be controlled by altering parameters like temperature, laser
intensity and chamber gas ambient chemistry and pressure.
ZnO nanorods are the primary structures of interest in this work. The reason
for this is because nanostructures, which have a small “footprint”, can be grown on a
variety of substrates with minimal strain in the main body of the nanorod, regardless
of lattice match at the interface, due to the free surfaces of the nanorod sidewalls,
which can relieve the effects of mismatch. Nanorods grown at high temperatures can
display excellent crystalline and optical properties, and can be grown with relatively
small amounts of source material and in short periods, enabling a range of systematic
experiments to be undertaken. The main growth method used here is a three step
process involving the deposition of a buffer layer of ZnO nanorods using drop
coating and CBD, followed by the primary growth of nanorods using CTR-VPT.
The growth method used here was developed in our group based upon other work in
the literature.60,61
Chapter 2 details this process. It is expanded and developed
further in this work to take account of the use of isotopically enriched growth
materials.
12
1.5 Defects in ZnO and the Isotope Effect
Low temperature photoluminescence (PL) is a powerful and useful technique
in the study of ZnO and other semiconductor materials. It can reveal information
about the electronic, vibrational and optical characteristics of the material under
investigation. It is therefore extremely useful in the study of defects and impurities
in sample because these phenomena induce changes in the electronic, vibrational
structure of the material, and therefore the optical characteristics also. Low
temperature PL is used extensively in this work to investigate defects and impurities
in ZnO. PL investigations are commonly carried out low temperature as this allows
much greater detail to be observed in the emission spectrum at high resolutions.
Typical temperatures for low temperature PL are < 20 K. PL relies on the
recombination of correlated electron-hole pairs called excitons following excitation
by photons with energy greater than the band gap. The excitation is typically carried
out using a laser beam with an above band gap photon energy. When the excitons
recombine, they emit a photon at an energy which is lower than that of the material’s
band gap. When these excitons are bound at the site of a defect or impurity, their
recombination gives rise to a photon at a particular energy (due to the specific
exciton confinement energy at the defect) which is unique to that defect or impurity,
and this allows PL to be used to study the defects in great detail. The growth method
used to produce ZnO nanorods in this work yields material with excellent optical
quality (in terms of both emission intensity and spectral line widths), which allows
high resolution PL investigations subsequently.
ZnO low temperature PL emission spectra typically display an intense
emission near the band gap energy in the UV which is called the band edge region.
This is mostly due to excitons bound to donors such as Al, Ga or In.88,89
ZnO also
often emits a broad band structure with specific spectral features in the visible range
called the structured green band (SGB), in addition to other broad defect-related
features in the visible region.7,90,91
This SGB emission has been assigned to a copper
(Cu) impurity atom with a zero phonon line (ZPL) at 2.86 eV92
and a broad band
multi-phonon emission at lower energies peaking at about 2.4 eV. This emission
was first posited to be related to Cu by Dingle in 1969.92
However, several other
sources, including native defects such as Zni and VO, have also been suggested as
13
being responsible for this emission.7,42,91
Cu was subsequently confirmed to be
involved in this defect by Byrne et al.93
One of the purposes of the present work is
to study the involvement of native defects in the centre responsible for this emission.
Very high optical quality is vital in such studies and nanostructured deposits can
provide the very high quality material required, as explained above.
Isotopic enrichment is another very useful technique in the study of defects in
semiconductor materials including ZnO, particularly in conjunction with optical
methods like PL. Briefly, the technique involves enriching the crystal with either
specific isotopes of an atom involved in, or suspected of involvement in, a defect or
enriching the surrounding native crystal lattice with specific isotopes. The
vibrational states in the crystal couple to electronic states, including those at defect
sites, to produce so-called vibronic levels. This is known as electron-phonon
coupling. Transitions between different vibronic levels, either at a defect or the
surrounding lattice produce the ZPL and phonon side-bands involving multiple
phonon replicas, as seen for the Cu-related green band in ZnO. Changes in the
energies of exciton recombinations may arise due to changes in both the isotopic
make up of a defect or a surrounding lattice, both of which affect the vibrational
nature of the vibronic levels because of the difference in mass of different
isotopes.94,95
PL, defects in semiconductors, the structure of the ZnO PL emission
spectrum, the Cu-related defect at 2.86 eV, and the isotope effect are all discussed in
more detail in chapter 3.
1.6 Motivation and Objectives
The first objective of this work was to adapt common previously used
methods of producing ZnO nanostructures such as nanorods to grow ZnO nanorods
which were enriched with particular isotopes of Zn. The natural Zn isotopic
abundance in ZnO is 48.6% 64
Zn, 27.9% 66
Zn, 4.1% 67
Zn, 18.8% 68
Zn and 0.6%
70Zn.
96 Natural ZnO therefore has an average atomic mass of
65.4ZnO. We set out to
synthesise ZnO nanorods enriched with 64
Zn, 66
Zn and 68
Zn. We have successfully
14
produced these Zn isotopically enriched ZnO nanorods by developing a novel
synthesis technique using CBD and VPT, based on previous work carried out in our
group. This new growth method is fast, relatively cheap and relatively easy to carry
out. Studies using isotopically enriched samples in the past have included work on
band gap energies, phonon positions and line widths and heat capacity, but in all
cases using bulk single crystal samples.96–99
The use of nanostructures, rather than
bulk single crystals, means that rather small amounts of isotopically enriched
material can be grown, which is advantageous given the cost of the isotopically
enriched source materials. It is also relevant to the potential applications of these
structures. The optical quality of these enriched samples, as observed using low
temperature PL, is excellent which makes them ideal for use in detailed optical
studies of defects and impurities.
The second objective was to grow ZnO material which was isotopically
enriched with 18
O isotopes. Natural O in its normal isotopic abundance state is
99.76% 16
O. This work proved more challenging, and ultimately ZnO nanorod
samples enriched with 18
O where successfully grown by two separate novel methods.
The first method was a modified VS VPT technique in Dublin City University
(DCU) based on the three step process developed by our group, and modified for the
Zn-enriched samples, further modified to produce O-enriched samples. The second
method was a VLS technique using the direct sublimation of ZnO powder at higher
temperatures. These growths were carried out in the University of Jena in Germany
with collaborators there. Details are given in chapter 2. The O-enriched samples
produced by both methods were of similar excellent optical quality to the Zn-
enriched samples, and therefore were also suitable for use in detailed optical studies
of defects and impurities using low temperature PL.
The third objective of this work was to carry out an investigation, using the
materials described above, into any possible involvement of native defects in the Cu-
related ZPL 2.86 eV emission and associated broad band SGB in ZnO. Isotopically
enriched material is very useful for this study since changing the isotopic content of
the material, either in the case of the Zn or O atoms, may result in anomalous shifts
(i.e. different to the overall shifts in band gap energy) in the energies of bound
exciton recombinations, as well as systemic changes in emission line widths. The
recombination energies of the band edge excitons, which occur near the band gap at
15
around 3.3 eV, and the 2.86 eV Cu-related emission, can be determined. If, in
addition to the known Cu atom, native defects such as Zni or O interstitials (Oi), or
Zn vacancies (VZn) or VO, are involved in the 2.86 eV emission, then the change in
band edge exciton recombination energies and line widths with isotope changes
might be different to the change in same for the 2.86 eV ZPL, due to the additional
presence of the native defects complexing with the Cu atom in this defect. The
effect of changing isotopic abundance is therefore used in this work to investigate
this defect. This study is carried out using low temperature PL and is made possible
by the very high optical quality of the isotopically enriched nanorods produced using
novel growth methods.
It is worth noting once again that this is the first time isotopically enriched
ZnO has been produced in nanostructured form, specifically nanorods, as key
previous studies of such material have used bulk single crystals.96–98,100
The optical
quality of these nanorods, in terms of their band edge emission intensities, and more
specifically their line widths, is significantly better than that reported previously for
ZnO bulk single crystals.96,97
This indicates their excellent structural quality and
their great potential for use in PL defect studies. These points constitute one of the
main original and novel features of the present work. Understanding the nature of
defects and impurities in semiconductor materials like ZnO is vital for the control
and optimisation of the optical and electronic properties of the materials and thus for
their potential use in many existing and proposed applications. As such this study
aims to achieve definitive conclusions on the involvement of native defects in the
Cu-related defect at 2.86 eV in ZnO. These experiments are described in chapters 4,
5 and 6.
Finally, a number of other experiments were carried out using ZnO samples
prepared during this work. This includes work carried out by the author solely, or in
collaboration with colleagues, both in DCU and elsewhere, using both natural and
isotopically enriched ZnO nanorods. This work demonstrates that the growth
methods and material produced here can be used in many different studies to provide
useful new information in a number of fields. These experiments are described in
appendix A.
16
1.7 Outline of Remainder of Thesis
This thesis describes the growth of isotopically enriched ZnO nanostructures
using novel, fast and relatively easy growth techniques. The morphology, crystal
structure and optical properties are studied. In particular, a comprehensive optical
study of the near band edge and Cu-related defect at 2.86 eV is carried out using low
temperature PL to determine any possible involvement of native defects in the 2.86
eV emission. A number of other applications and experiments using these methods
and materials are then described. A brief outline of the remainder of the thesis is as
follows:
Chapter 2 describes in detail the experimental techniques used throughout
this work. This includes the growth of ZnO nanorods and Zn-isotopically enriched
ZnO nanorods using a novel modified combined CBD and VPT technique and two
growth techniques used to produce the O-isotopically enriched ZnO samples in DCU
and in Jena. This chapter also describes the experimental setups of the investigative
techniques used to study the structural and optical characteristics of samples
throughout this work.
Chapter 3 presents background theory relevant to the PL optical studies on
defects in this work in detail. This includes background theory on the PL technique
itself, as well as general information on defects in semiconductors, a description of
the typical structure of the PL emission spectrum in ZnO and specific background
information on the Cu-related defect at 2.86 eV in ZnO. This chapter also discusses
the effect of isotope substitution on the PL spectrum and its relevance to this work.
Chapter 4 describes the results of the growth of Zn isotopically enriched ZnO
nanorods. This includes studies on the morphology, crystal quality, isotopic
enrichment and optical quality using a number of characterisation techniques.
Chapter 5 presents the results of the growth of O isotopically enriched ZnO
nanorods using the two separate methods used in DCU and in Jena. The results of
studies on the morphology, crystal quality, isotope enrichment and optical quality are
presented.
17
Chapter 6 looks at the optical properties of isotopically enriched ZnO. This
includes studies on the shifts of the band edge exciton recombination energies as
well as shifts in the 2.86 eV ZPL with changing isotope enrichments using low
temperature PL spectroscopy. This chapter also discusses what these results mean in
terms of involvement of native defects such as interstitials or vacancies in the Cu-
related defect at 2.86 eV.
Chapter 7 outlines the final general conclusions of the thesis and includes
some suggestions for further work on the topics in the thesis.
1.8 References
1 A.K. Biswas, Indian J. Hist. Sci. 28, 309 (1993).
2 C. Jagadish and S. Pearton, editors , Zinc Oxide Bulk, Thin Films and
Nanostructures (Elsevier, Oxford, 2006).
3 C.F. Klingshirn, B.K. Meyer, A. Waag, A. Hoffmann, and J. Geurts, ZnO: From
Fundamental Properties Towards Novel Applications (Springer-Verlag, Berlin,
Heidelberg, 2010).
4 D. Reynolds, D. Look, B. Jogai, C. Litton, T. Collins, W. Harsch, and G. Cantwell,
Phys. Rev. B 57, 12151 (1998).
5 S.J. Xu, W. Liu, and M.F. Li, Appl. Phys. Lett. 81, 2959 (2002).
6 M. Willander, O. Nur, Q.X. Zhao, L.L. Yang, M. Lorenz, B.Q. Cao, J. Zúñiga
Pérez, C. Czekalla, G. Zimmermann, M. Grundmann, A. Bakin, A. Behrends, M. Al-
Suleiman, A. El-Shaer, A. Che Mofor, B. Postels, A. Waag, N. Boukos, A. Travlos,
H.S. Kwack, J. Guinard, and D. Le Si Dang, Nanotechnology 20, 332001 (2009).
7 U. Ozgur, Y.I. Alivov, C. Liu, a. Teke, M. a. Reshchikov, S. Dogan, V. Avrutin,
S.-J. Cho, and H. Morkoc, J. Appl. Phys. 98, 041301 (2005).
8 C.J. Lee, T.J. Lee, S.C. Lyu, Y. Zhang, H. Ruh, and H.J. Lee, Appl. Phys. Lett. 81,
3648 (2002).
9 M.H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo, and
P. Yang, Science (80-. ). 292, 1897 (2001).
18
10 Y.-H. Chou, B.-T. Chou, C.-K. Chiang, Y.-Y. Lai, C.-T. Yang, H. Li, T.-R. Lin,
C.-C. Lin, H.-C. Kuo, S.-C. Wang, and T.-C. Lu, ACS Nano 9, 3978 (2015).
11 F.D. Nayeri, E.A. Soleimani, and F. Salehi, Renew. Energy 60, 246 (2013).
12 M. Law, L.E. Greene, J.C. Johnson, R. Saykally, and P. Yang, Nat. Mater. 4, 455
(2005).
13 C. Opoku, K.F. Hoettges, M.P. Hughes, V. Stolojan, S.R.P. Silva, and M.
Shkunov, Nanotechnology 24, 405203 (2013).
14 H. Nanto, T. Minami, and S. Takata, J. Appl. Phys. 16, 482 (1986).
15 J.X. Wang, X.W. Sun, Y. Yang, H. Huang, Y.C. Lee, O.K. Tan, and L.
Vayssieres, Nanotechnology 17, 4995 (2006).
16 B. Ul Haq, R. Ahmed, A. Shaari, A. Afaq, B. a. Tahir, and R. Khenata, Mater. Sci.
Semicond. Process. 29, 256 (2015).
17 A. Mesquita, F.P. Rhodes, R.T. da Silva, P.P. Neves, A.O. de Zevallos, M.R.B.
Andreeta, M.M. de Lima, A. Cantarero, I.S. da Silva, M. a. Boselli, X. Gratens, V.A.
Chitta, A.C. Doriguetto, W.B. Ferraz, A.C.S. Sabioni, and H.B. de Carvalho, J.
Alloys Compd. 637, 407 (2015).
18 Y.Q. Fu, J.K. Luo, X.Y. Du, A.J. Flewitt, Y. Li, G.H. Markx, A.J. Walton, and
W.I. Milne, Sensors Actuators B Chem. 143, 606 (2010).
19 Z.L. Wang and J. Song, Science (80-. ). 312, 242 (2006).
20 S.H. Jo, J.Y. Lao, Z.F. Ren, R.A. Farrer, T. Baldacchini, and J.T. Fourkas, Appl.
Phys. Lett. 83, 4821 (2003).
21 E. McCarthy, S. Garry, D. Byrne, E. McGlynn, and J.-P. Mosnier, J. Appl. Phys.
110, 124324 (2011).
22 S. Garry, É. McCarthy, J.-P. Mosnier, and E. McGlynn, Nanotechnology 25,
135604 (2014).
23 X. Huang, X. Wang, S. Wang, J. Yang, L. Zhong, and J. Pan, Nanoscale 5, 5596
(2013).
24 K. Memarzadeh, A.S. Sharili, J. Huang, S.C.F. Rawlinson, and R.P. Allaker, J.
Biomed. Mater. Res. Part A 103, 981 (2015).
19
25 Wurtzite Crystal Structure - Wikipedia, the free Encyclopedia, [Online]. Accessed
13-02-2015. Available at: http://en.wikipedia.org/wiki/Wurtzite_crystal_structure
26 K. Govender, D.S. Boyle, P.B. Kenway, and P. O’Brien, J. Mater. Chem. 14, 2575
(2004).
27 A. Ashrafi and C. Jagadish, J. Appl. Phys. 102, 071101 (2007).
28 H. Liu, J.S. Tse, and H. Mao, J. Appl. Phys. 100, 093509 (2006).
29 E.H. Kisi and M.M. Elcombe, Acta Crystallogr. Sect. C Cryst. Struct. Commun.
45, 1867 (1989).
30 J.E. Jaffe and A.C. Hess, Phys. Rev. B 48, 7903 (1993).
31 H. Sowa and H. Ahsbahs, J. Appl. Crystallogr. 39, 169 (2006).
32 S. Desgreniers, Phys. Rev. B 58, 14102 (1998).
33 D. Byrne, The Growth and Characterisation of Ordered Arrays of Zinc Oxide
Nanostructures and Optical Studies of Defects in Zinc Oxide, Dublin City
University, 2012.
34 R. Aggarwal, H. Zhou, C. Jin, J. Narayan, and R.J. Narayan, J. Appl. Phys. 107,
113530 (2010).
35 N. Fujimura, T. Nishihara, S. Goto, J. Xu, and T. Ito, J. Cryst. Growth 130, 269
(1993).
36 R. Yousefi and B. Kamaluddin, J. Alloys Compd. 479, L11 (2009).
37 C. Klingshirn, Phys. Status Solidi 244, 3027 (2007).
38 P. Schröer, P. Krüger, and J. Pollmann, Phys. Rev. B 47, 6971 (1993).
39 R.T. Girard, O. Tjernberg, G. Chiaia, S. Söderholm, U.O. Karlsson, C. Wigren, H.
Nylén, and I. Lindau, Surf. Sci. 373, 409 (1997).
40 S.-H. Wei and A. Zunger, Phys. Rev. B 37, 8958 (1988).
41 S.B. Zhang, S.-H. Wei, and A. Zunger, Phys. Rev. B 52, 13975 (1995).
42 M.D. McCluskey and S.J. Jokela, J. Appl. Phys. 106, 071101 (2009).
43 D.C. Look, D.C. Reynolds, J.R. Sizelove, R.L. Jones, C.W. Litton, G. Cantwell,
and W.C. Harsch, Solid State Commun. 105, 399 (1998).
44 A. Janotti and C.G. Van de Walle, Phys. Rev. B 76, 165202 (2007).
20
45 C.G. Van de Walle, Phys. B Condens. Matter 308-310, 899 (2001).
46 C.G. Van de Walle, D.B. Laks, G.F. Neumark, and S.T. Pantelides, Phys. Rev. B
47, 9425 (1993).
47 L. Yu, Q. Zhu, D. Fan, and Z. Lan, J. Wuhan Univ. Technol. Mater. Sci. Ed. 27,
1184 (2012).
48 Y. Marfaing and A. Lusson, Superlattices Microstruct. 38, 385 (2005).
49 D.C. Look, D.C. Reynolds, C.W. Litton, R.L. Jones, D.B. Eason, and G. Cantwell,
Appl. Phys. Lett. 81, 1830 (2002).
50 M. Joseph, H. Tabata, and T. Kawai, Jpn. J. Appl. Phys. 38, L1205 (1999).
51 K. Kim, D.-H. Lee, S. Lee, G.-E. Jang, and J.-S. Kim, Nanoscale Res. Lett. 7, 273
(2012).
52 M. Kumar, T.-H. Kim, S.-S. Kim, and B.-T. Lee, Appl. Phys. Lett. 89, 112103
(2006).
53 Y.. Ryu, S. Zhu, D.C. Look, J.M. Wrobel, H.M. Jeong, and H.W. White, J. Cryst.
Growth 216, 330 (2000).
54 Y.I. Alivov, E. V Kalinina, a E. Cherenkov, D.C. Look, B.M. Ataev, a K.
Omaev, M. V. Chukichev, and D.M. Bagnall, Appl. Phys. Lett. 83, 4719 (2003).
55 J.B. Kim, D. Byun, S.Y. Ie, D.H. Park, W.K. Choi, J.-W. Choi, and B. Angadi,
Semicond. Sci. Technol. 23, 095004 (2008).
56 S. Vempati, S. Chirakkara, J. Mitra, P. Dawson, K. Kar Nanda, and S.B.
Krupanidhi, Appl. Phys. Lett. 100, 162104 (2012).
57 X. Zhang, F. Herklotz, E. Hieckmann, J. Weber, and P. Schmidt, J. Vac. Sci.
Technol. A Vacuum, Surfaces, Film. 29, 03A107 (2011).
58 H. Xu, H. Wang, Y. Zhang, W. He, M. Zhu, B. Wang, and H. Yan, Ceram. Int. 30,
93 (2004).
59 M. Mikami, T. Eto, J. Wang, Y. Masa, and M. Isshiki, J. Cryst. Growth 276, 389
(2005).
60 L.E. Greene, M. Law, D.H. Tan, M. Montano, J. Goldberger, G. Somorjai, and P.
Yang, Nano Lett. 5, 1231 (2005).
61 D. Byrne, E. McGlynn, K. Kumar, M. Biswas, M.O. Henry, and G. Hughes, Cryst.
21
Growth Des. 10, 2400 (2010).
62 J. Grabowska, a. Meaney, K. Nanda, J.-P. Mosnier, M. Henry, J.-R. Duclère, and
E. McGlynn, Phys. Rev. B 71, 115439 (2005).
63 T. Long, S. Yin, K. Takabatake, P. Zhnag, and T. Sato, Nanoscale Res. Lett. 4,
247 (2009).
64 X.Y. Kong and Z.L. Wang, Nano Lett. 3, 1625 (2003).
65 Y. Wang, X. Chen, J. Zhang, Z. Sun, Y. Li, K. Zhang, and B. Yang, Colloids
Surfaces A Physicochem. Eng. Asp. 329, 184 (2008).
66 S. Vempati, A. Shetty, P. Dawson, K. Nanda, and S.B. Krupanidhi, J. Cryst.
Growth 343, 7 (2012).
67 S. Vempati, J. Mitra, and P. Dawson, Nanoscale Res. Lett. 7, 470 (2012).
68 D. Bao, H. Gu, and A. Kuang, Thin Solid Films 312, 37 (1997).
69 R.B. Peterson, C.L. Fields, and B.A. Gregg, Langmuir 20, 5114 (2004).
70 M. Afzaal, M. a. Malik, and P. O’Brien, New J. Chem. 31, 2029 (2007).
71 S.-H. Yi, S.-K. Choi, J.-M. Jang, J.-A. Kim, and W.-G. Jung, J. Colloid Interface
Sci. 313, 705 (2007).
72 K. Kakiuchi, E. Hosono, T. Kimura, H. Imai, and S. Fujihara, J. Sol-Gel Sci.
Technol. 39, 63 (2006).
73 M. Kokotov and G. Hodes, J. Mater. Chem. 19, 3847 (2009).
74 M.N.R. Ashfold, R.P. Doherty, N.G. Ndifor-Angwafor, D.J. Riley, and Y. Sun,
Thin Solid Films 515, 8679 (2007).
75 M. Wang, E.J. Kim, S.H. Hahn, C. Park, and K. Koo, Cryst. Growth Des. 8, 501
(2008).
76 J. Grabowska, K.K. Nanda, E. McGlynn, J.-P. Mosnier, and M.O. Henry, Surf.
Coatings Technol. 200, 1093 (2005).
77 M. Biswas, E. McGlynn, M.O. Henry, M. McCann, and a. Rafferty, J. Appl. Phys.
105, 094306 (2009).
78 C. Borchers, S. Mu, D. Stichtenoth, D. Schwen, and C. Ronning, J. Phys. Chem. B
110, 1656 (2006).
22
79 D. Hou, T. Voss, C. Ronning, A. Menzel, and M. Zacharias, J. Appl. Phys. 115,
233516 (2014).
80 D. Byrne, R. Fath Allah, T. Ben, D. Gonzalez Robledo, B. Twamley, M.O. Henry,
and E. McGlynn, Cryst. Growth Des. 11, 5378 (2011).
81 R. Triboulet and J. Perrière, Prog. Cryst. Growth Charact. Mater. 47, 65 (2003).
82 D.J. Park, D.C. Kim, J.Y. Lee, and H.K. Cho, Nanotechnology 17, 5238 (2006).
83 J.R. Arthur, Surf. Sci. 500, 189 (2002).
84 Y.W. Heo, M. Kaufman, K. Pruessner, D.P. Norton, F. Ren, M.F. Chisholm, and
P.H. Fleming, Solid. State. Electron. 47, 2269 (2003).
85 B. Doggett, S. Chakrabarti, R. O’Haire, a. Meaney, E. McGlynn, M.O. Henry,
and J.P. Mosnier, Superlattices Microstruct. 42, 74 (2007).
86 Y. Sun, G.M. Fuge, and M.N.R. Ashfold, Superlattices Microstruct. 39, 33 (2006).
87 A. Fouchet, W. Prellier, B. Mercey, L. Mechin, V.N. Kulkarni, and T. Venkatesan,
J. Appl. Phys. 96, 3228 (2004).
88 M. Schilling, R. Helbig, and G. Pensl, J. Lumin. 33, 201 (1985).
89 S. Muller, D. Stichtenoth, M. Uhrmacher, H. Hofsass, C. Ronning, and J. Roder,
Appl. Phys. Lett. 90, 012107 (2007).
90 J. Čížek, J. Valenta, P. Hruška, O. Melikhova, I. Procházka, M. Novotný, and J.
Bulíř, Appl. Phys. Lett. 106, 251902 (2015).
91 H. Morkoç and U. Özgür, Zinc Oxide Fundamentals, Materials and Device
Technology (Wiley-VCH, Weinheim, 2009).
92 R. Dingle, Phys. Rev. Lett. 23, 579 (1969).
93 D. Byrne, F. Herklotz, M.O. Henry, and E. McGlynn, J. Physics. Condens. Matter
24, 215802 (2012).
94 E. McGlynn, A Photoluminesence Study of Cadmium and Aluminium-Related
Defects in Silicon, Dublin CIty University, 1996.
95 B. Henderson and G.F. Imbusch, Optical Spectroscopy of Inorganic Solids
(Clarendon Press, Oxford, 1989).
96 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
23
Cardona, Solid State Commun. 128, 35 (2003).
97 S. Tsoi, X. Lu, A.K. Ramdas, H. Alawadhi, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 74, 165203 (2006).
98 J. Serrano, F. Manjón, A. Romero, F. Widulle, R. Lauck, and M. Cardona, Phys.
Rev. Lett. 90, 055510 (2003).
99 J. Serrano, R. Kremer, M. Cardona, G. Siegle, a. Romero, and R. Lauck, Phys.
Rev. B 73, 094303 (2006).
100 J. Serrano, F. Widulle, A.H. Romero, A. Rubio, R. Lauck, and M. Cardona, Phys.
Status Solidi 235, 260 (2003).
24
Chapter 2: Growth and
Characterisation Techniques
2.1 Introduction
This chapter presents the experimental techniques used during this work. It
describes both the growth techniques used to synthesise the ZnO nanorod samples,
including isotopically enriched ZnO, and the characterisation techniques used to
examine and study the material subsequently. Section 2.2 describes the basic growth
procedure for ZnO nanorods, and how it was modified to produce Zn isotopically
enriched samples. The growth procedure is based upon a three-step process
previously developed previously in our group.1 This involves (i) the deposition of a
seed layer of ZnO, (ii) growth of a ZnO buffer layer by CBD, and finally (iii) the
growth of c-axis aligned ZnO nanorods by VPT. Section 2.3 describes the growth
of the O isotopically enriched ZnO samples by two separate methods, one in DCU
and another in the University of Jena in Germany with collaborators there. The
characterisation techniques are then described in section 2.4. The morphology and
crystal quality of the samples were examined using scanning electron microscopy
(SEM) and X-ray diffraction (XRD). The isotopic enrichment was examined using
secondary ion mass spectroscopy (SIMS). Some energy dispersive x-ray
spectroscopy (EDX) was also performed. The optical quality was examined using
Raman spectroscopy, and the primary optical technique in this work, that of low
25
temperature PL spectroscopy. Optical reflectance studies were also carried out on
some samples. This chapter describes the experimental details of these techniques as
well as some brief background information on them. Some further information on
the growth and optical techniques are described in subsequent chapters, alongside
their respective results.
2.2 Growth of Zn Isotope-enriched Nanorods
This section describes the basic procedures for the growth of ZnO nanorods as
developed by our group. The procedure is then modified to produce Zn-enriched
isotopically pure nanorods in this work.
2.2.1 Substrate Preparation
The large majority of ZnO nanorods in this work were grown on Si
substrates. Typically phosphorous-doped n-type Si with a (100) surface orientation
was used. Some growths were also performed on other substrates such as fused
quartz, a-plane sapphire (Al2O3) or c-plane sapphire and the growth methods used
produced very similar deposits on such substrates. Si was chosen as the substrate as
it is relatively cheap and easily available. It is also electrically conductive which is
useful in characterisation techniques like SEM and SIMS. Typically, Si (100)
wafers were cleaved into small pieces of approximately 1-4 cm2 in size using a
diamond scribe. The pieces of Si were then blown with a stream of nitrogen (N) to
remove any dust or other particles from the surface. The substrate was then cleaned
before use by sonication in acetone for approximately 10 minutes, rinsing with fresh
acetone, sonication in ethanol for a further 10 minutes and then rinsing in fresh
ethanol. They were then dried in a nitrogen stream. No attempt was made to remove
the native silicon dioxide (SiO2) layer.
26
2.2.2 Seed Layer Preparation
The first step in the growth of the ZnO is the deposition of a thin seed layer
of ZnO on the Si substrate. Si is not epitaxially lattice matched with ZnO therefore a
seed layer of ZnO is required to provide nucleation sites for subsequent aligned
nanorod growth. This seed layer was deposited on the bare Si substrate using the
method reported by Greene and Law et al.2,3
A 0.005 M solution of Zn acetate
dihydrate (Riedel-de Haen) in absolute ethanol was prepared. This solution was
sonicated for up to about an hour until the Zn acetate had completely dissolved in the
ethanol and was then left to cool before use. The seed layer was deposited by drop
coating. Approximately 3.75 µl of solution per cm2 of sample area was applied to
the substrate surface using a pipette. This droplet was allowed to spread so that it
evenly coated the entire surface and was left for 20 seconds before being rinsed off
the surface with copious amounts of fresh ethanol. The substrate was then dried in a
gentle nitrogen stream. This process was repeated four more times for each sample.
During this process, H2O from the atmosphere diffuses into the solution and reacts
with the Zn acetate to form Zn hydroxide precipitates which settle on the substrate.
It has been reported that the ambient humidity is a factor affecting this process and
the number of seed layers can need to be varied to account for it.4 However this is
not something which was found to be an issue during this work. The substrates were
then annealed at 350˚C in air for 30 minutes to decompose the Zn hydroxide into
ZnO. This process produces a thin layer of crystallographically aligned ZnO
crystallites on the surface which act as nucleation sites for nanorod growth at later
stages.5,6
2.2.3 Chemical Bath Deposition
The second phase of the three step growth process is the deposition of a
buffer layer of aligned ZnO nanorods on the Si substrate by CBD. Three different
chemical baths were used during this work, although one was used a large majority
of the time. This primary chemical bath used was an aqueous solution of NaOH and
Zn nitrate.1,7
This reaction produces a metastable Zn(OH)4-2
solution and the ZnO is
27
deposited from this. The second bath used was an aqueous solution of Zn nitrate and
HMT. The third chemical bath was a simple solution of Zn acetate in deionised
water (DI-H2O). This results in the direct thermal decomposition of the Zn acetate
into Zn hydroxide followed by ZnO. In each case, the solutions were heated for
varying period of time during deposition. The laboratory equipment used for this
step is similar for all three chemical bath techniques. Figure 2.1 shows a schematic
and a photograph of the equipment used. The reactions took place in aqueous
solutions in a beaker using a hotplate as the heat source. The seeded substrates were
suspended in the reaction solutions as shown. Polytetrafluoroethylene (PTFE)-
coated magnetic stirring bars were used to stir the solutions, while the temperature of
the solutions was monitored using either an alcohol thermometer or a digital
temperature probe. Each of the three chemical baths is described in turn below.
(a) (b)
Figure 2.1: (a) Schematic diagram, and (b) photograph, of the experimental setup
used for the CBD growth step.
28
(i) NaOH-based reaction
The NaOH-based reaction is the primary chemical bath used during this work
and a large majority of samples were produced using this method. Before this step,
for the NaOH chemical bath, it is important that the equipment used is clean and free
from residual ZnO from any previous growths. Residual ZnO could result in the
formation of unwanted precipitates in the solution. The beakers, stirring bars and
spatulas to be used were cleaned using a dilute solution of sulphuric acid (H2SO4)
and then rinsed with copious amounts of DI-H2O and allowed to dry. The H2SO4
reacts with any ZnO to form a soluble zinc sulphate solution which can be easily
removed by rinsing with copious amounts of DI-H2O, leaving a clean beaker. A
solution of 0.02 M Zn nitrate hexahydrate (99.998%, Alfa Aesar) in DI-H2O was
prepared and vigorously stirred until the Zn nitrate was fully dissolved. A separate
solution of 0.8 M NaOH (99%, VWR) was prepared and was also vigorously stirred
until the NaOH was fully dissolved. The Zn nitrate solution was then slowly added
to the NaOH solution while stirring vigorously. It was important to stir the mixture
vigorously while adding the Zn nitrate to disperse it throughout the mixture and
prevent any precipitation of Zn hydroxide from the solution. The solution was clear
and does not contain any precipitates when prepared correctly. The mixture was
heated to approximately 70 °C and stirred gently. The sample was then submerged
in the solution for approximately 25 minutes while the temperature was maintained
at 70 °C and the solution stirred gently. The sample was then removed, washed with
copious amounts of DI-H2O and dried with a gentle nitrogen stream. Typically a
total volume of 160 ml was used for the bath. This was prepared by mixing 80 ml
each of Zn nitrate and NaOH solution to give a total reaction volume of 160 ml with
concentrations of 0.01 M Zn nitrate and 0.4 M NaOH. This volume of solution
allowed for a number of substrates to be submerged at once, usually four or five.
However the total volume can be altered (usually to 40 ml or 80 ml) as needed for
smaller numbers of samples. The molarities were kept constant when this was done.
This bath leaves a layer of densely packed c-axis aligned ZnO nanorods which acts
as an effective buffer layer for subsequent growth of larger nanorods using VPT.
29
(ii) HMT-based reaction
To prepare this bath, HMT (99.5%, Fluka) and Zn nitrate hexahydrate (99.998%,
Alfa Aesar) were dissolved in DI-H2O in concentrations of 25 mM. The Zn nitrate
solution was prepared first and dissolved by stirring. HMT was then added to the Zn
nitrate solution and dissolved by stirring. The ZnO seeded substrate was then
submerged in the bath and the solution was heated to 90 °C for 30 minutes. The
sample was then removed, washed with DI-H2O and dried in a gentle nitrogen
stream. The volume used in this case was typically 80 ml, although this can be
varied as needed. An equimolar 25 mM solution equates to 7.4 mg Zn nitrate and
3.5 mg HMT per ml of reaction volume. This reaction produces significant amounts
of precipitates in the solution, which can deposit on the substrate surface. This
reaction was therefore rarely used in this work.
(iii) Acetate-based reaction
Zn acetate dihydrate (Riedel-de Haen) was dissolved by stirring in DI-H2O in a
concentration of 25 mM. As with the other baths, the volume of solution can be
varied with typical volumes of 40 ml or 80 ml being used. The seeded substrates
were submerged in the solution and then heated to about 65 °C. This temperature
was maintained for up to three hours, but sometimes less, with continuous stirring.
Precipitates also form in the bath, so the substrate would be removed, washed with
DI-H2O, and placed in a freshly prepared bath half way through a three hour growth.
Following growth, the sample was washed with DI-H2O and dried in a gentle
nitrogen stream. The bath was also used rarely during the course of this work.
30
2.2.4 Vapour Phase Transport
The third stage in the growth of ZnO nanorods is CTR-VPT using a VS
technique. ZnO will not readily deposit on bare SiO2. In addition, ZnO and Si are
not lattice matched so any deposition (including by methods other than VPT) will
generally result in non-aligned nanorods, especially with the presence of a native
oxide layer on the Si. The buffer layer of crystallographically textured ZnO
nanorods previously deposited by CBD thus carries out two important roles: it
provides energetically favourable sites for the nucleation of ZnO and it allows for the
growth vertically aligned c-axis nanorods, following the buffer layer alignment. The
VPT step was carried out in a fused quartz tube with an internal diameter of 37 mm.
The tube was placed inside a large alumina tube, which was itself inside a Lenton
Thermal Designs single temperature zone horizontal tube furnace. The tube was
connected to a supply of high purity Argon (99.999%) controlled by a mass flow
controller (MFC, model Analyt GFC 17). In a typical growth of unenriched nat
ZnO,
60 mg of high purity ZnO powder (99.9995%, Alfa Aesar) and 60 mg of graphite
powder (99.9999%, Alfa Aesar) were carefully weighed out using a mass balance.
The two powders were carefully mixed together and placed in a mortar. This
mixture was then ground together with a pestle in order to increase the contact area
between the two powders and produce a fine homogeneous powder mixture. This
powder was then carefully spread over an area of about 2 cm in the middle of a small
alumina boat. The sample was suspended above the powder, using thin strips of Si
and the edge of the boat as supports, with the ZnO buffer layer facing downwards
towards the powder. The alumina boat was then placed into the tube so that the
sample was in the centre of the furnace. The quartz tube was sealed and purged with
an Ar flow of 90 sccm for about 5-10 minutes. (There was an exhaust at the opposite
end of the tube to the Ar MFC.) The temperature was then increased to the desired
level of around 850 °C – 950 °C for 1 hour with the Ar flow remaining at 90 sccm.
Most of the growths carried out were at a temperature of 925 °C. After one hour, the
furnace was allowed to cool. When the temperature reached about 350 °C, the Ar
flow was stopped and the alumina boat and sample removed. Figure 2.2 shows a
schematic diagram and photograph of the furnace apparatus during the VPT step.
31
(a)
(b)
Figure 2.2: (a) Schematic diagram of the furnace setup used for the VPT growth of
ZnO nanorods, and (b) photograph of the VPT furnace.
The three step growth procedure above called for 60 mg ZnO and 60 mg
graphite powders for the VPT stage. However this proved impracticable for the
growth of isotopically enriched ZnO with specific Zn isotope enrichments. This was
because the isotopically enriched material is quite expensive and only relatively
small amounts of Zn-enriched ZnO powders were available. Therefore the amount
of ZnO and graphite powders was reduced to 10 mg of each for the growth of the
Zn-enriched nanorods. Nanorods were successfully grown with this lower amount of
source powder, in both natural and isotopically pure states, and without any other
changes to the procedure. This was expected as only a small amount of the source
32
powders was generally vapourised during the normal (i.e. 60 mg + 60 mg) VPT
growth process, with the rest remaining in the alumina boat.
nat
ZnO contains Zn isotopes in the proportions shown in table 2.1.8 In this
work, isotopically enriched ZnO samples were grown by substituting the ZnO
powder with ZnO powders (Isoflex) enriched with 64
Zn, 66
Zn and 68
Zn. Samples of
64ZnO,
66ZnO and
68ZnO nanorods were therefore grown. The Zn enrichment levels
of each source powder is shown table 2.1. Samples were also grown with different
proportions of different Zn isotopes present. Samples with equal proportions of two
different isotopes were grown using 5 mg of each powder along with 10 mg of
graphite powder (64/66
ZnO, 64/68
ZnO and 66/68
ZnO). Similarly, a sample of three
different Zn isotopes was produced using 3 mg of each powder along with 10 mg of
graphite (64/66/68
ZnO). A separate quartz tube, alumina boat, mortar and pestle were
used for each sample. The Zn-enriched samples had the usual natural oxygen
isotopic distribution of 99.76% 16
O.
natZn isotopic content
Zn isotopic content in
enriched source powders
64Zn 48.6% 99.94%
66Zn 27.9% 99.29%
67Zn 4.1%
68Zn 18.8% 99.34%
70Zn 0.6%
Table 2.1: Natural abundances of Zn isotopes in ZnO and the enrichment levels of
the source powders used to produce Zn-isotopically enriched nanorods.
2.3 Growth of O Isotope-enriched ZnO Nanorods
Following on from the growth of Zn-enriched ZnO nanorods, this section
describes the growth of O-enriched nanorods. This work proved to be more
33
challenging. However, after extensive trials, which are further outlined in chapter 5,
O-enriched ZnO nanorods were successfully grown by two separate methods. The
first was a VS method carried out by modifying the CTR-VPT setup used in the
growth of Zn-enriched samples. The second was a VLS technique carried out in
collaboration with colleagues in the Institute for Solid State Physics in the University
of Jena, Germany. These growths were carried out by Mr. Lukas Trefflich and the
author in the group of Prof. Carsten Ronning during a research visit by the author in
September 2015.
2.3.1 Method 1: Modified VS CTR-VPT method in DCU
The CTR-VPT growth described in section 2.2.4 is based on the reduction of
ZnO powder by the graphite to produce Zn vapour and carbon monoxide (CO). The
Zn vapour is then re-oxidised in a VS process at the energetically favourable sites
provided by the aligned CBD buffer layer using residual O2 present in the tube
following the Ar flush. In order to grow ZnO nanorods enriched with 18
O isotopes, it
was necessary to remove all the residual O2 from the tube, and then reintroduce 18
O2
gas. A schematic of the setup for this experiment is shown in figure 2.3.
Figure 2.3: Schematic diagram of the experimental setup for the growth of O-
enriched ZnO nanorods using a modified VS technique.
34
The furnace was loaded as before with an alumina boat containing the mixed
powder of 60 mg graphite and 60 mg ZnO in its natural isotopic state with a Si
substrate previously deposited with a buffer layer of ZnO nanorods by CBD
suspended above the powder. The exhaust end of the tube was then sealed using a
valve. At the other end, a series of valves connect the tube to a vacuum pump, N and
O gases and Ar gas via the MFC as shown. A small regulator was used to control
the pressure of the 18
O2 (99%, Sigma Aldrich) lecture bottle. A digital pressure
gauge was also present. The tube was slowly evacuated using the vacuum pump to a
pressure of <1 mbar. Following this, the tube was refilled to atmospheric pressure
with artificial ‘air’, that is, with ~21% 18
O2 and ~79% N2. This mixture was left for
about 15 minutes to allow the gasses to mix and fill the entire tube evenly. The Ar
flush at 90 sccm was then started, the exhaust valve opened, and after 5 minutes the
temperature was raised to 925°C for 1 hour as before.
Three samples were produced using this method: natural Zn16
O by
evacuating the tube and re-filling with 16
O2 and N2, isotopically enriched ZnO by re-
filling with 18
O2 and N2, and a mixed 50:50 Zn16/18
O sample by using a mixture of
both oxygen isotopes with N2.
2.3.2 Method 2: VLS VPT method in Jena
O-enriched ZnO nanorods were also produced using a VLS method with an Au
catalyst in Jena. The method used was based on previous work by that group.9,10
The setup shows some similarities to that described in section 2.3.1 albeit also with
some notable differences. The furnace in Jena contained two alumina tubes, one
inside the other. Approximately 120 mg of ZnO powder acted as the source and was
spread over about 5-6 cm in an alumina boat and placed at the centre of the furnace.
About six to eight pieces of Si substrate had 10 nm of Au deposited on them by
plasma sputter coating and were placed in another alumina boat about 16 cm from
the ZnO powder, over a range of a few cm. The tube was sealed and evacuated to a
pressure of <0.3 mbar. The temperature was raised to 1350°C over a period of 5
hours with a pre-determined ramp controlled by a computer (400°C for 90 mins,
1150°C for 180 mins, 1350°C for 60 mins). When the temperature reached ~600°C
35
the Ar (99.999%) was introduced at 50 sccm in the reverse direction, i.e. moving any
vapour away from the substrates. The pressure was held constant at 100 mbar from
this point on. After the temperature was at 1350°C for 60 minutes, the growth then
took place as the Ar direction was reversed for a further period of 60 minutes and
carried the Zn and O2 vapours from the source powder to the substrates where they
deposited. In this VLS method, the Au coating melts and forms droplets on the
substrate which act as energetically favourable nucleation sites. The nanorods in this
case are not preferentially aligned. The furnace then cooled overnight before that
samples were removed. Figure 2.4 shows the basic setup of this system.
Figure 2.4: Schematic diagram showing the general experimental setup for the
growth of O-enriched ZnO nanorods using the VLS technique in Jena.
In this case it is the ZnO source powder which provides both the Zn and O2
vapours for subsequent growth due to the sublimation of the ZnO powder. In order
to produce O-enriched ZnO nanorods therefore, it was necessary to produce enriched
ZnO source powder. Zn powder (99.9%, Alfa Aesar) was oxidised in DCU by
placing ~160 mg of such powder in an alumina boat in the furnace and evacuating
the tube to < 1 mbar. The tube was then filled with ~200 mbar of either 16
O2 or 18
O2
and ~550 mbar of N2 giving a pressure of ~ 750 mbar. This mixture was left to settle
for about 15 minutes before the furnace was heated to 800°C for one hour. About
60-90 mg of oxidised powder was recovered from the boat after each run due to
36
some of the produced ZnO depositing on the tube edges. Separate boats and tubes
were used for each isotope and smaller tubes of 18 mm internal diameter were used
here to reduce the amount of 18
O2 gas needed as it was in limited supply.
Three growths were then carried out in Jena using these oxidised Zn16
O and
Zn18
O powders: Zn16
O, Zn18
O and mixed 50:50 Zn16/18
O using 60 mg of each
powder.
2.4 Characterisation Techniques
This section describes the experimental tools use to characterise the samples
produced during this work. The primary method of initial characterisation was the
use of SEM to examine the morphology of CBD and VPT grown nanorods. A
number of other techniques were then used to study the structural, isotopic, chemical
and optical properties. These include XRD, SIMS, EDX, Raman spectroscopy, low
temperature PL spectroscopy and reflectance spectroscopy. In some cases external
collaborations were leveraged to carry out aspects of these characterisation
techniques. Where this occurred details are presented in the relevant sections
including who performed the experiments and where they took place.
2.4.1 Scanning Electron Microscopy
SEM was used throughout this work to examine the morphology of the
samples. It was necessary to use SEM rather than optical microscopy due to the
scale of the features under investigation. Under the Rayleigh criterion, two point
sources are considered to be just resolved when the central maximum of the
diffraction pattern of one point overlaps with the first minimum of the second point’s
diffraction pattern, yielding a resolution limit which is linearly proportional to the
light wavelength. Since the features under examination here are close to, and in
many cases smaller, in size than the wavelength of visible light, an optical
microscope was not suitable for examining the morphology of these samples.
37
Imaging with beams of electrons provides a higher resolution as needed in this work.
This is possible since the electrons have a much smaller de Broglie wavelength than
visible wavelengths and thus the same diffraction limitations are not present.
SEM works by directing a beam of high energy electrons at the sample. A
schematic of a typical SEM system is shown in figure 2.5(a) and (b). The electron
beam originates from the electron gun which contains a source such as a tungsten or
lanthanum hexaboride cathode. Current is passed through the source and the
electrons are emitted by thermionic emission. This is an example of a hot cathode.
Cold cathodes can also act as electron sources through field emission. An anode
provides an accelerating voltage to the emitted electrons. The beam travels through
a column under vacuum, and is shaped, directed and focussed as required by a
system of apertures and magnetic lenses by using the Lorentz force. These magnetic
lenses can shape and direct the electron beam in a similar way to how optical
refracting lenses alter a beam of light. The condenser lens adjusts the spot size and
beam current and the objective lens adjusts the beam focus. The beam is raster
scanned over the sample surface rather than incident over a large area
simultaneously. The sample chamber is also under vacuum (~ 10-3
mbar). The
sample chamber is brought to atmospheric pressure when changing the sample.
The electrons interact with the target material in a number of different ways
which produce a number of different signals by which information can be gathered
about the sample under investigation. Electrons deflected away from the sample
surface in an elastic scattering process are called backscattered electrons. This
process depends strongly on the mass of the atoms in the material, with higher
atomic masses leading to greater numbers of backscattered electrons. These
electrons therefore contain information about the material composition of a sample,
with image contrast indicating average elemental atomic weight. Most of these
electrons travel back up the optic axis so a circular scintillation or solid state detector
surrounding the beam emission point collects these electrons.
When the incident electrons collide with an outer shell electron in the sample
they can cause the ejection of these electrons from the sample’s atoms. These
ejected electrons are called secondary electrons following this inelastic scattering
process. Secondary electrons generally have very low energies (the definition is
38
electrons with energies less than 50 eV) so only ones close to the surface tend to
escape the sample. There is a contrast change as the angle between the sample
normal and the incident beam increases. This is due to more secondary electrons
being emitted as the exposed surface area increases at higher angles. This contributes
to the high quality of image formed using these electrons and can be used to
determine the orientation of structures on the surface. They are therefore used to
collect topographical information about the sample surface. This information is
again displayed in the image by changes in contrast. Low energy secondary
electrons are attracted by a positive bias to a scintillation detector and multiplied by a
photomultiplier tube to produce the signal used to form the image. SEM images
have a large depth of field and can display large areas of the sample surface in some
detail. Secondary electron imaging was used to produce the images in this work as
the focus of this characterisation technique for our purposes was on morphology.
During the inelastic interaction that creates the secondary electrons, another
useful signal is also produced. Some ejected electrons can leave behind an empty
state in the atom’s inner shell. This hole can then combine with another electron
from the outer orbitals of that same atom. The energy difference between the
orbitals is then released upon recombination in the form of x-rays. This is illustrated
in figure 2.5(c). These x-rays can be detected and they are characteristic of the
element they came from and therefore can be used to obtain elemental information
about a sample or map differences in composite materials. This is the signal used in
EDX. This technique has been used occasionally in this work, particularly during
early stages of the work on the growth of O-enriched nanorods.
In this work, most SEM characterisation was carried out using a Carl-Zeiss
EVO series SEM system with a lanthanum hexaboride source and fitted with
secondary electron and EDX detectors. A photograph of this system is shown in
figure 2.5(d). The system can also be used in backscatter mode and variable pressure
mode, which can be useful for electrically insulating or biological samples. SEM
was used to examine the coverage of nanorods on the surface of samples before they
were examined using low temperature PL. It was also of interest to observe the
orientation and shape of the nanorods as this can vary slightly between growth runs.
39
(a) (b)
(c) (d)
Figure 2.5: (a) Schematic diagram showing the main components of a typical SEM
system and (b) a more detailed diagram showing the secondary electron (SE) and
backscattered electron (BSE) detectors and the associated electronics (from
reference 11). A camera was used instead of the CRT in this work to view the
sample. (c) Diagram of the process of generation of EDX signal, (d) Photograph of
the Carl-Zeiss SEM system in DCU.
40
Some SEM images in this work, specifically the images of the O-enriched
samples grown in the University of Jena, were obtained there using a FEI Helios
NanoLab 600i system by Mr. Lukas Trefflich with the author.
2.4.2 X-ray Diffraction
XRD works on the principles of elastic scattering and constructive
interference. When a coherent beam of electrons is incident on a sample, some of
the x-rays are inelastically scattered resulting in an x-ray of longer wavelength.
These waves will not interfere with one another as they will have difference phases
and wavelengths to each other. This is known as Compton scattering. They are not
relevant here. However, most of the incoming x-rays will be scattered elastically
from the sample. This is known as Thompson scattering. The physical structure of a
crystalline sample can be viewed as a series of crystal planes of atoms along
different directions. The spacing and orientation of these planes is rooted in the
crystal structure of the particular material. Elastic scattering and diffraction occurs
as x-rays bounce off different layers of atomic planes in the crystal. In some cases,
the scattered x-rays from different planes can constructively interfere with each other
under certain conditions as they have a definite phase relationship. The condition for
constructive interference to occur between all the x-rays scattering from difference
planes is described by the Bragg equation (Eqn. 2.1).12,13
2𝑑𝑠𝑖𝑛𝜃 = 𝑛𝜆 Eqn. 2.1
where d is the distance between atomic planes in the sample, θ is the incident angle
for the incoming x-rays, λ is the incident x-ray wavelength and n is an integer
indicating the order. The left hand side of the Bragg equation, 2dsinθ, is the
additional path length travelled by an x-ray scattering off adjacent planes. When this
is equal to a whole number of wavelengths constructive interference occurs. Figure
2.6 shows a diagram of the geometry of this situation. Constructive interference of
41
x-rays coming from different planes occurs at a particular incident angle and results
in a sharp peak in the x-ray diffraction spectrum at that angle. Otherwise there is
destructive interference between the scattered x-rays and no peak is observed.
(a)
(b)
Figure 2.6: (a) Schematic diagram of the layout of an XRD system with zoom
showing diffraction from adjacent crystal planes and the geometry of the Bragg
equation, and (b) photograph of the Bruker AXS D8 Advance Texture
Diffractometer.
42
The general geometry of an XRD system is shown in figure 2.6(a), along
with a photograph of the Bruker AXS D8 Advance Texture Diffractometer used in
this work in figure 2.6(b). The incident angle is changed by rotating the sample
through the angle θ. The diffraction angle is moved through the angle θ by moving
the detector through the angle 2θ, so that the situation is that of symmetrical
reflection with respect to the sample (i.e. equal incident and reflected angles at all
times). The x-ray source does not move. This method is used to collect the 2θ-ω
spectra, where ω=θ. The detector collects the strong x-ray peaks at the angles when
the Bragg equation is satisfied. The other major method of operation that was used
in this work is called the rocking curve. In this case, the detector is placed at the
correct 2θ angle of a previously observed peak corresponding to a particular plane or
feature. The sample is then rotated about the angle ω. The rocking curve generally
gives a peak that can be used to compare the crystal quality and degree of
nanocrystal alignment between samples.
XRD can be used, with the Bragg equation, to determine the orientation of
the crystal planes in a sample and the spacing d between those planes. The peak
positions of many materials are contained in databases (for example JCPDS14
) which
can be consulted in order to identify materials or contaminants in a sample based on
the peaks recorded. It is therefore an important technique for material identification
and characterisation.
XRD was primarily used throughout this work to confirm the presence of c-
axis aligned ZnO as well as to obtain a quantitative measure of the crystal quality
and orientation consistency, i.e. to study the crystalline quality of the deposited
material.
2.4.3 Secondary Ion Mass Spectroscopy
SIMS is a mass spectroscopic technique used to analyse material from
samples, including nanostructures and thin films. An ion beam is directed at the
sample at a high voltage. This is called the primary ion beam. Upon impact, the ion
beam sputters the surface and material is ejected from the sample. Most of the
43
material ejected is not charged, but a small amount is ejected as ions. These
secondary ions are then attracted to a detection system by a positive or negative bias,
depending on the charge on the ions under investigation. The detection system is
designed to differentiate between different materials based on their mass by
measuring their charge/mass ratios. In this way different atoms or molecules can be
detected. Of specific relevance for this work, isotopic distributions can be examined.
SIMS was used in this work to examine isotopic distributions in enriched ZnO
samples. The measurements take place under vacuum to ensure the maximum
amount of sputtered material reaches the detector and contamination is minimised.
Figure 2.7(a) shows a schematic of the main components in a SIMS system.
Although SIMS is used to examine surfaces (the primary beam typically sputters to a
depth of a few nm), depth profiling is also a very useful SIMS technique. In this
case the primary beam continuously sputters the sample and ‘digs’ down past the
first few atomic layers. This produces a depth profile of different layers in a
composite material. SIMS can scan over a large area, making 3D visualisations
possible, as well as examining a single point on a sample surface. When the ion
current is set relatively low in order to scan surfaces, it is referred to as static SIMS.
By contrast, dynamic SIMS uses a high current density and a focussed beam to
maximise sputtering yields and examine bulk material. This is the mode of
operation used in depth profiling.
Two different systems were used to carry out SIMS in this work. In the first
SIMS system (labelled simply as ‘SIMS’), measurements were performed in an
ultra-high vacuum system with a base pressure of 1x10-9
mbar. The setup included a
Hiden Analytical IG20 ion gun operating with a positive Ar (99.99%) ion beam of 5
keV at a chamber partial pressure of 5x10-5
mbar with the analysis chamber not
exceeding 2x10-8
mbar during analysis. The sample current during analysis ranged
from 90 nA to 110 nA. All samples were analysed at an angle of ~20° off normal
incidence. This was to try to reduce any effects from the unenriched ZnO buffer
layers below the isotopically pure VPT-grown nanorods. Secondary Zn ions were
detected using a Hiden Analytical mass spectrometer EQS quadrupole analyser.
This is a positive mode SIMS setup i.e. it detects positively charged ions. The actual
measurements on this system were carried out by Mr. Conor Byrne in DCU. A
photograph of this system is shown in figure 2.7(b)
44
(a) (b)
(c)
Figure 2.7: (a) Schematic diagram showing the sputtering process used in SIMS, (b)
photograph of the SIMS system and (c) photograph of the Millbrook miniSIMS alpha
system.
The second system used to carry out SIMS measurements was the benchtop
Millbrook miniSIMS alpha system (labelled as ‘miniSIMS’ to distinguish it from the
first system). A photograph of this system is shown in figure 2.7(c). This uses a
positive Ga ion beam at 6 keV and a quadrupole analyser. The samples were
mounted at a 45° angle in this case. The miniSIMS alpha is a complete SIMS
system designed for ease of use and quick changing of samples. The sample stage
can be ejected, samples mounted and the stage returned to vacuum in about half an
hour. The stage also held several samples at once which again speeds up the process
of loading and scanning. The system operates at a pressure of about 1x10-7
mbar.
45
All of the systems components, including the Ga source, the sample stage, the
detection system and the vacuum systems are contained inside one benchtop
apparatus. The PC software allows full control of the entire system. The miniSIMS
is capable of switching between positive and negative mode, depending on the ions
under investigation, by changing the polarity of the detection system. The
instrument is also fitted with charge neutralisation system which allows insulating
samples to be examined. Measurements on this system were carried out by the
author.
2.4.4 Raman Spectroscopy
Raman spectroscopy is an optical technique used to examine vibrational,
rotational and other low-frequency excitation modes in a material by detecting shifts
in energy of scattered light. When a monochromatic light source, usually in the form
of a laser, is incident upon a material, it can interact with that material in a number of
ways. The simplest form of interaction is for the photons to be scattered elastically.
This is known as Rayleigh scattering. Raman scattering occurs when the scattered
photon has interacted with phonon modes in the sample and therefore has an energy
that is different to that of the incident photon, be it more or less. This is an inelastic
process and is much less frequent than Rayleigh scattering with only around one
photon in a million undergoing Raman scattering.
Briefly, the incident photon excites the target atom or molecule from some
vibrational or rotational state into a higher virtual state. The scattered photon is then
emitted but the atom or molecule may end up in a different excitation state to where
it began. If the final state has more energy than the original state then it follows that
the emitted photon must have less energy than the incident photon. This is called a
Stokes shift. Likewise, if the final state has an energy which is less than the original
state, then the photon emitted must have a greater energy than the incident photon.
This is called an anti-Stokes shift. Figure 2.8 shows a diagram of each of these
interactions. Raman scattering ultimately comes about due to the strong electric
field associated with the incoming laser beam interacting with the electrons in the
target material and inducing a dipole moment in those atoms or molecules, in
46
combination with the modulation of the dipole polarizability at the excitation
frequency of e.g. the vibration or rotation. The combination of these effects leads to
the inelastic scattering (essentially sidebands in a classical picture).
Figure 2.8: Energy level diagram showing Rayleigh scattering, and Raman Stokes
and anti-Stokes scattering processes.
The vibrational and rotational mode energies of an atom or molecule are
unique and Raman can therefore be used to detect and study these properties in
specific atoms or molecules and utilise the information as a “fingerprint” for specific
materials. Similarly, bulk solids, liquids and gases also have characteristic phonon
modes. In this work, shifts in the Raman lines indicate shifts in the characteristic
phonon modes in isotopically enriched ZnO material due to the changes in atomic
mass.
A typical Raman system will include a laser source, efficient collection and
focussing optics, relevant filtering to remove the much stronger elastically scattered
signal and a charge-coupled device (CCD) or photomultiplier tube (PMT) detection
system. Raman measurements on the 64
ZnO, 66
ZnO and 68
ZnO samples were
performed by Dr. Joseph Cullen at Linköping University, Sweden. They were
performed at room temperature using a Horiba LabRAM micro-Raman system.
Non-resonant excitation was used to make Raman measurements in the
47
( , ) zz x y x y back-scattering geometry using the 532 nm line of a solid state laser
with z-direction coinciding with the direction of the c-axis of ZnO. This essentially
means that the laser was incident on the sample normally along the z-direction and c-
axis of the crystal, and that the scattered photons back along this same direction were
collected for analysis. Resonant excitation, where the incident photon energy is
tuned to an electronic transition in the sample to enhance the associated vibrational
modes, is not used here. Raman measurements on the O-enriched samples were
obtained in DCU with a similar setup using a Jobin Yvon LabRam HR800 system at
room temperature using an Ar laser at 488 nm and an air-cooled CCD detector, with
the assistance of Dr. Rajani Vijayaraghavan.
2.4.5 Photoluminescence
The optical properties of the ZnO nanorod samples were analysed using low
temperature PL. The details of the experimental setups and their operation are
presented here. More detail on the theory of PL and its application in the study of
defects and impurities in semiconductors, and in particular ZnO, is presented in
chapter 3. Two separate low temperature PL systems were used in this work. They
both have advantages, and they are both described in detail below. In short, samples
were placed in a Janis Research closed cycle helium (He) cryostat and cooled to
temperatures in the 10-20 K range, depending on the system. The samples were
excited by directing a continuous wave helium cadmium (HeCd) laser at a
wavelength of 325 nm onto them. The emitted PL light was then collected using a
lens and directed to the entrance aperture of the spectrometer. The two systems
utilised in the collection of the PL emission from the samples were (i) a diffraction
grating spectrometer system and (ii) a Fourier transform (FT) spectrometer system.
48
(i) Diffraction grating spectrometer system (SPEX)
(a)
(b)
Figure 2.9: (a) Schematic, and (b) photograph, of the optical setup used for PL with
the SPEX monochromator.
The dispersion spectroscopic system is based on a 1 m model SPEX 1704
monochromator. The optical setup is shown in figure 2.9. The sample was placed in
a Janis Research model SHI-950-5 cryostat and cooled to ~12 K. The HeCd laser
beam was directed onto the sample by a simple system of mirrors. The sample
49
surface was at an angle of 45° to the laser beam. The emitted light was carefully
collected by a focussing lens (at 90° to the incident laser beam) and focussed onto
the entrance slit of the SPEX monochromator. Light entering the monochromator
via the variable width entrance slit was directed onto the diffraction grating and then
to the detector via the variable width exit slit. A mercury (Hg) spectral lamp was
placed in the region of the cryostat so that the light emitted also fell in the entrance
slit. The spectral lines from this lamp were used to calibrate the spectra recorded to
correct for minor variations from scan to scan due to minor hysteresis in the
mechanical drive system which moves the grating. The 365.0158 nm Hg line in each
spectrum was aligned to this position. This is the position of this line in air.15
In
addition, the SPEX spectra have been corrected for the refractive index of air by
dividing by 1.000285.16
The spectra were recorded using software on a PC. This
dispersion-based system using a diffraction grating essentially scans through the
wavelengths by having the grating direct each one in turn onto the exit slit and into
the detector. The intensity at each wavelength is then recorded. Schott glass filters
were placed at the laser (UG11) to remove longer wavelength lines and at the
monochromator entrance slit (WG-345) to remove any shorter wavelength laser and
plasma tube light.
The detector was a Hamamatsu model R3310-02 PMT in photon counting
mode which was cooled to approximately -20 °C using a Peltier system EMI
FACT50 Cooler. The dark count of the detector at this temperature is very low at
around 30 counts per second or about 5 nA.17
The monochromator grating was
blazed at 330 nm with 1200 grooves/mm (ISA model 510-05). The efficiency of this
system is limited by two main factors. These are the efficiency of the diffraction
grating, and the efficiency of the PMT, both of which vary with wavelength. The
grating is very efficient around the band edge region in ZnO of ~369 nm, which is a
region of strong interest in this work. It is less efficient at the green band energies
of around 430-600 nm. The detector is also less efficient at longer wavelengths.
Equations and their graphs representing the efficiency curves for the PMT and
diffraction grating used in this system can be found in appendix A, as reproduced
from reference 18. Because of this some samples needed to be annealed at 900˚C for
10 minutes in order to increase the SGB intensity when we wished to study that
relatively weak emission using this system. In addition, some of the O-enriched
50
samples were annealed in this way to activate the SGB emission as it was not
observed in the as-grown samples.19
This is further elaborated on in chapter 3. The
resolving power, R, of the diffraction grating is related to the resolution, Δλ, by the
formula
𝑅 =𝜆
Δ𝜆= 𝑁𝑚 Eqn. 2.2
where N is number of grooves in the grating, m is the diffraction order and λ is the
wavelength.18,20
The system’s f number is ~f/9 and the lenses and optics are chosen
to match this. The maximum resolution of this system is 0.008 nm (~0.07 meV).
Typically the band edge spectra were recorded using a step size of 0.005 nm (with a
resolution of ~0.07 meV) and a slit width of 6 µm, and the SGB ZPLs with a step
size of 0.01 nm (~0.09 meV) and a slit width of 20 µm. These slit widths are at or
below typical pixel sizes in CCD cameras commonly used to record spectra in
monochromator systems. Additionally, regarding the calibration, although the Hg
lamp was placed slightly off the optical axis of the system and therefore one might
expect some broadening of these lines as not all of the light would fall on the
diffraction grating, the observed lines remained very sharp and there was sufficient
light to use the Hg lines for calibration purposes.
(ii) Fourier transform spectrometer system (FT)
The second spectroscopic system consisted of a Bomem DA8 FT
spectrometer and IPH8200L PMT detector. The optical setup is shown in figure
2.10. The sample was placed in a Janis Research CSS-550 cryostat and cooled to
~19 K. The 325 nm HeCd laser beam was directed onto the sample using mirrors
and a focussing lens to produce a small, intense spot on the sample. The sample was
again at 45° to the incident beam. The emitted light was collected by a lens (at 90°
to the incident laser beam) which produced a parallel beam and focussed onto the
entrance aperture by a parabolic mirror. In this case the entrance aperture was a
variable circular aperture. There was another variable circular aperture at the exit
point just before the detector. The spectra were again collected using a PC.
51
(a)
(b)
Figure 2.10: (a) Schematic, and (b) photograph, of the optical setup used for PL
with the FT spectrometer.
The FT spectrometer works on a fundamentally different principle to the
diffraction-based dispersion monochromator. Many of the factors that can limit the
resolution of a typical dispersion system, such as the slit widths and the grating
dispersion, are either not as important or not applicable to an FT system. The FT
system is based around the Michelson interferometer. Light entering the system is
52
split into two beams. Each beam travels along different paths and are then
recombined and incident on the detector. One beam travels along a fixed arm and
reflects off a stationary mirror. The path length of the second beam is variable by
the movement of the mirror from which it reflects. When the two beams recombine
they interfere with each other. With light of a single frequency, if the path difference
travelled by the two beams equals a whole number of wavelengths, then constructive
interference occurs. Otherwise, different interference conditions occur including
completely destructive interference when the path difference travelled by the beams
corresponds to a difference of half a wavelength. As the moving mirror moves, the
detector sees a series of bright and dark fringes in a sinusoidal pattern resulting from
the interference of the two beams with changing phase difference. The intensity
detected by the PMT is therefore related to the phase difference between the beams
and the frequency of the light. However, in examining a spectrum from a sample
using PL the signal will contain many difference frequencies of different intensities.
The intensity recorded at the detector as the mirror moves will be a complex pattern
of interference fringes containing the interference patterns of each individual
frequency present in the beam and their intensities. This is called an interferogram.
The interferogram is the Fourier transform of the intensity spectrum. Therefore, the
spectrometer software performs an inverse Fourier transform on the interferogram,
extracts the frequencies and intensities contained therein and produces the optical
spectrum.
To briefly describe the relationship between the interferogram and the
spectrum we consider the case of a monochromatic source and the interaction in the
Michelson interferometer after beam splitting of two standard electromagnetic waves
with a phase difference of φ:
𝐸1 = 𝐸 cos(𝑘𝑥 − 𝜔𝑡) 𝑎𝑛𝑑 𝐸2 = 𝐸 cos(𝑘𝑥 − 𝜔𝑡 + 𝜑) Eqn. 2.3
where k is the wavenumber which is related to the wavelength λ by the relation
2π / λ, and ω is the angular velocity. x and t represent position and time respectively.
The detector output is the intensity of the wave incident on it, rather than the electric
field. The total intensity IT is given by the time average of the square of the total
electric field ET as follows:
𝐼𝑇 = ⟨𝐸𝑇2⟩ = ⟨(𝐸1 + 𝐸2)2⟩ = ⟨𝐸1
2 + 𝐸22 + 2𝐸1𝐸2⟩ Eqn. 2.4
53
This can then be written as
𝐼𝑇 = ⟨𝐸𝑇2⟩ =
𝐸2
2+
𝐸2
2+ 𝐸2 cos 𝜑 Eqn. 2.5
𝐼𝑇 = ⟨𝐸𝑇2⟩ = 𝐸2 + 𝐸2 cos 𝜑 Eqn. 2.6
𝐼𝑇 = ⟨𝐸𝑇2⟩ = 𝐸2(1 + cos 𝜑) Eqn. 2.7
The phase difference between the waves φ can be written as
𝜑 =2𝜋𝐷
𝜆= 𝑘𝐷 Eqn. 2.8
where D is the difference in distance travelled by the two waves. Equations 2.7 and
2.8 show that the intensity at the detector varies in a cosinusoidal pattern as a
function of the difference in the distance travelled by the two waves (related to the
moving mirror) and the wavelength of the light. This relation between the intensity
and the mirror position is what produces the interferogram recorded at the detector.
However the spectra recorded using an FT spectrometer never contains only one
wavelength because even a very close to monochromatic source will not produce
light of a single wavelength. There will always be a spread of wavelengths
described by the line width. A full spectrum will contain signal at many
wavelengths and at different intensities. All these interferograms combined are
observed at the detector. In order to capture all of this information, we introduce a
function I(k) which represents the entire spectrum under investigation. Integrating
equation 2.7 gives the total intensity detected as
𝐼(𝐷) = ∫ 𝐼(𝑘)(1 + cos(𝑘𝐷)) 𝑑𝑘∞
−∞ Eqn. 2.9
This interferogram is the Fourier transform of the emission spectrum under
investigation. It contains all the wavelength and intensity information needed to
reconstruct the original spectrum. If we perform an inverse Fourier transform we
obtain the original spectrum:
𝐼(𝑘) = ∫ 𝐼(𝐷)(1 + cos(𝑘𝐷)) 𝑑𝐷∞
−∞ Eqn. 2.10
This is computed by the PC software. The limits on these integrals are
infinity, but in practice the system is limited by the range of the moving mirror. The
54
longer the distance available for it to move, the higher the resolution the system will
have. If we consider a single wavelength, the mirror can only move a certain
distance, P. Essentially the integral stops at P, and the effect is a broadening
inversely proportional to P, with the achievable resolution limited to 1/P. The
minimum resolvable Δk for the instrument (representing the line width of an ideal
monochromatic source) is given by21
Δ𝑘 ≅𝜋
𝑃 Eqn. 2.11
This can also be written in units of cm-1
as22
Δ𝜐 ≈0.7
2𝑃 Eqn. 2.12
For example, if the mirror travelled a distance of 3 cm, then the value of Δν would
be ~0.12 cm-1
. The system performance is also limited by a number of factors which
complicate the brief description presented here, including aperture size, mirror
alignment, detector noise, optical imperfections and the process of converting the
signal from analogue to digital. The system contains a HeNe laser which it uses to
align the fixed mirror to correct for any misalignment during a scan. The instrument
must also determine the zero path difference (ZPD) position. A broad band white
light source is used for this. At the ZPD point a large peak occurs as all the
wavelengths from this source undergo constructive interference allowing the ZPD
point to be determined to high accuracy.
The FT system has a number of distinct advantages in comparison to
dispersive grating spectrometer systems. The system records all wavelengths at once
as they are all contained in the interferogram recorded. Dispersive monochromators
must scan through each wavelength individually. This is known as the multiplex
advantage. This can speed up the time it takes to obtain a spectrum. This also
produces a signal-to-noise advantage of √N, for N spectral channels, excluding
photon noise. The circular aperture can allow a much higher throughput which can
be important when examining weak signals. This is known as the throughput
advantage. It reflects the fact that FT systems do not need their entrance slits to be
as narrow as in dispersive systems to produce the same resolution. The resolving
power, R, of a Michelson interferometer due to the aperture is given by23
55
𝑅 =2𝜋
Ω Eqn. 2.13
with
Ω = 𝜋 (𝜃
2)
2
Eqn. 2.14
where θ is the solid angle of admittance at the entrance aperture. There is thus a
trade-off between higher resolutions and higher signal strength.
However this system still suffers from signal-to-noise problems when
measuring very weak signals, as the dark count of the detector was higher than the
SPEX system, and therefore a range of filters were used to eliminate spectral regions
not of interest when carrying out different scans. Filtering out parts of the spectrum
that were not of interest also helps to stop a weak signal from being lost amongst an
intense signal. The white light source had all wavelengths below 700 nm filtered
out. In addition, a second filter in front of the PMT further reduced interference
from wavelengths not of interest during a particular scan. A BG25 Schott glass
bandpass filter was used (cut on ~330 nm, cut off ~480 nm) for scans in the UV
region, and a 606HSP visible bandpass filter (cut on ~380 nm, cut off ~620 nm) was
used for the visible region. A further bandpass filter (3RD, cut on ~430 nm, cut off
~440 nm) was used in conjunction with the 606HSP filter when measuring the ZPL
line at 2.68 eV as it very weak. A filter (UG11) was also placed in front of the laser
as with SPEX system to remove longer wavelength laser plasma lines. The software
also has the option of running many scans and adding them together to reduce the
overall signal-to-noise ratio. This was normally done during this work. The FT
system is also very accurate and stable, and did not need any external calibration as
used with the SPEX system to correct for any minor shifts in wavelength. The
system also accounted for the refractive index of air automatically. A detailed
description of issues related to Fourier transforms can be found in reference 22. On
this system the band edge spectra were recorded with a resolution of a 0.015 meV for
the band edge and 0.15 meV for the SGB ZPLs.
56
2.4.6 Reflectance Spectroscopy
(a)
(b)
Figure 2.11: (a) Schematic, and (b) photograph, of the optical setup used during
reflectance spectroscopy experiments with the FT spectrometer.
Reflectance spectroscopy was performed by using a modified optical set up
with the FT spectrometer. This is shown in figure 2.11. A broad band deuterium
57
lamp was used as the light source. This was collimated using a lens and then
focussed to a small spot on the sample. The sample was again at 45° to the incident
light. Reflected light was collected by another lens (at 90° to the incident light
beam) to form a parallel beam before being directed using parabolic mirror to the
entrance slit of the spectrometer. These spectra were collected with a pre-recorded
reference spectrum collected using a reflective piece of Al-coated Si automatically
subtracted by the software. Reflectance spectroscopy aims to record light reflected
from the surface of the samples, rather than light generated by photoluminescence
from the samples. It is a useful technique that can be used to record the position of
the free excitons and this was the purpose of using this characterisation technique in
this work.24
2.5 References
1 D. Byrne, E. McGlynn, K. Kumar, M. Biswas, M.O. Henry, and G. Hughes, Cryst.
Growth Des. 10, 2400 (2010).
2 L.E. Greene, M. Law, D.H. Tan, M. Montano, J. Goldberger, G. Somorjai, and P.
Yang, Nano Lett. 5, 1231 (2005).
3 M. Law, L.E. Greene, J.C. Johnson, R. Saykally, and P. Yang, Nat. Mater. 4, 455
(2005).
4 Y.-J. Lee, T.L. Sounart, D.A. Scrymgeour, J.A. Voigt, and J.W.P. Hsu, J. Cryst.
Growth 304, 80 (2007).
5 R.B. Saunders, E. McGlynn, M. Biswas, and M.O. Henry, Thin Solid Films 518,
4578 (2010).
6 D. Byrne, R. Fath Allah, T. Ben, D. Gonzalez Robledo, B. Twamley, M.O. Henry,
and E. McGlynn, Cryst. Growth Des. 11, 5378 (2011).
7 R.B. Peterson, C.L. Fields, and B.A. Gregg, Langmuir 20, 5114 (2004).
8 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
Cardona, Solid State Commun. 128, 35 (2003).
58
9 C. Borchers, S. Mu, D. Stichtenoth, D. Schwen, and C. Ronning, J. Phys. Chem. B
110, 1656 (2006).
10 D. Hou, T. Voss, C. Ronning, A. Menzel, and M. Zacharias, J. Appl. Phys. 115,
233516 (2014).
11 L. Reimer, Scanning Electron Microscopy: Physics of Image Formation and
Microanalysis, 2nd ed. (Springer-Verlag, Berlin, Heidelberg, New York, 1998).
12 J.J. Rousseau, Basic Crystallography (John Wiley & Sons, Sussex, 1998).
13 B.D. Cullity and S.R. Stock, Elements of X-Ray Diffraction, 3rd ed. (Prentice Hall,
New Jersey, 2001).
14 International Centre for Diffraction Data. Joint Commitee on Powder Diffraction
Standards (JCPDS) Database. Available at: http://www.icdd.com/
15 Kramida, A. Ralchenko, Yu. Reader, J. NIST ASD Team. NIST Atomic Spectra
Database (ver. 5.1), [Online]. Accessed 29-10-2013. Available at:
http://physics.nist.gov/asd
16 Stone, J.A. Zimmerman, J.H. Engineering Metrology Toolbox, Refractive Index
of Air Calculator, [Online]. Accessed 07-11-2013. Available at
http://emtoolbox.nist.gov/Wavelength/Edlen.asp
17Hamamatsu R3310-02 Photomultiplier Tube Specif. Sheet (n.d.).
18 D. Gorman, Photoluminescence and Excitation Studies of Semiconductors, Dublin
City University, 2001.
19 N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, and G. Cantwell,
Appl. Phys. Lett. 81, 622 (2002).
20 J. Cullen, Photoluminescence Studies of ZnO Doped with Stable and Radioactive
Impurities, Dublin City University, 2013.
21 M. Biswas, Growth and Characterisation of ZnO Nanostructures: Excitonic
Properties and Morphology, Dublin City Univeristy, 2010.
22 J. Chamberlain, The Principles of Interferometric Spectoscopy (Wiley, Chichester,
59
1979).
23 J. Fryar, Optical and AFM Studies of ZnO: Excitonic Properties, Surface
Morphology and Etching Effects, Dublin City Univeristy, 2005.
24 C.F. Klingshirn, B.K. Meyer, A. Waag, A. Hoffmann, and J. Geurts, ZnO: From
Fundamental Properties Towards Novel Applications (Springer-Verlag, Berlin,
Heidelberg, 2010).
60
Chapter 3: Background Theory of
Elements of Optical Properties of
ZnO
3.1 Introduction
This chapter presents the background information regarding relevant aspects of
the optical properties of ZnO under investigation in this work. We begin by
introducing the technique of low temperature PL which is a powerful technique in
the study of defects and impurities in semiconductor materials. Some general
information about defects and impurities in semiconductors is then presented,
followed by a general description of the main features of the ZnO PL emission
spectrum. One of these features, the Cu-related defect emission at 2.86 eV, is then
discussed in detail. This defect is of particular interest in this work. The use of the
isotope effect in semiconductor defect studies, and how it pertains to this work, is
described. Finally, the main phonon modes investigated using Raman spectroscopy
in this work, are described.
61
3.2 Low Temperature Photoluminescence
PL is a powerful and useful technique in the study of semiconductor materials
such as ZnO. It can reveal information about the electronic, vibrational and optical
characteristics of a material. It is therefore of exceptional use in the study of defects
and impurities in materials as these introduce distinctive changes in the electronic
and vibrational structure of the material. It is also a non-destructive and non-contact
technique which is an obvious advantage in its use.
PL relies on the creation of electrons and holes in the material and the analysis
of the photons released during the recombination of same. Electrons are excited
from the valence band to the conduction band when a light source with photon
energy greater than the band gap is incident on the sample (the ZnO band gap varies
from 3.37 K at 300 K to about 3.44 eV at low temperatures1,2
). Typically the source
of light is a laser. The excited electron leaves a positive hole in the valence band.
The electron and hole have a greater energy than the band gap due to the energy
imparted to it by the incident photon which was above the band gap energy of the
material. These electrons and holes are called hot carriers. They undergo a process
called thermalisation where they lose this excess energy to the surrounding crystal,
e.g. by collisions with phonons etc. The electron is then localised in energy close to
the edge of the conduction band and the hole localised in energy close to the valence
band edge (we neglect electron-hole Coulombic interactions for the present and
return to this issue later). They can then recombine in a radiative process and emit a
photon of energy equal to the band gap of the material. The thermalisation process
occurs at a much shorter timescale (~5 ps) than the recombination (>200 ps) and
therefore an excess population of electrons and holes accumulate at the band edges
before recombination.
We must also consider whether the material is a direct or indirect band gap
semiconductor. ZnO is a direct band gap material. This means that the crystal
momentum, or wave vector, k, is the same for position of the energy at the top of the
valence band and the bottom of the conduction band. In this case the electron and
hole can recombine in a radiative process involving only a photon since overall
momentum can be conserved because the photon has a very small momentum. The
62
energy, and therefore the wavelength, of the photon emitted during recombination is
equal to the band gap of the material and energy is thus conserved. This process
involves two particles, the electron and the photon, and therefore the quantum
mechanical matrix element, or probability, of this transition is much higher than non-
radiative processes. This contributes strongly to direct band gap materials like ZnO
being very efficient optical emitters and absorbers.
Figure 3.1: Diagram showing the excitation and recombination processes in (a)
direct, and (b) indirect, band gap semiconductors.
However, in an indirect band gap material, the edges of the valence and
conduction bands do not have the same crystal momentum values. In this case a
change in crystal momentum is required for recombination to occur, which cannot be
supplied solely by a photon because of the small momentum of the photon. This
generally happens by the emission or absorption of phonons to or from the crystal
lattice. The emitted photon can therefore be of greater or lesser energy than the band
gap of the material, although at room temperatures (and below) the density of
phonons is generally so low that the photon and phonon emission process dominates
and the photon therefore has a lower energy than the band gap. Due to the
63
involvement of this additional particle in the process, the quantum mechanical
probability of this occurring is much lower than the probability of recombination in a
direct band gap material. Competing non-radiative processes therefore can dominate
and thus it follows that indirect band gap materials are much less efficient optical
emitters and absorbers. The excitation, thermalisation and emission (by direct or
indirect band gap) processes described above are shown schematically in figure 3.1.
When an electron-hole pair is created by excitation as previously described,
they experience a Coulombic attraction to each other due to the electron’s negative
charge and the hole’s effective positive charge. The electron and hole can become
coupled to each other and move through the crystal around their centre of mass as a
single ‘quasi-particle’ which is called an exciton. When excitons travel without
restriction through the crystal (in very perfect and pure crystals with no defects or
impurities) they are called free excitons (FX).3,4
They can be thought of in a way
similar to the Bohr model of the atom and their binding energy and radius can
therefore be calculated using this model taking account of the electron and hole
effective masses and the materials dielectric permeability. They have a formation
energy slightly smaller than the band gap due to the Coulombic binding energy (60
meV in ZnO).3
Two types of excitons are generally identified for different materials.3 The
first are called Frenkel excitons. These are generally present in insulators with very
small dielectric constants. In this situation, the material’s ability to screen the
electric field due to the electron and hole charges is weak and so the Coulombic
attraction between the electron and the hole is very strong. These excitons are
tightly bound and have small radii of the order of one unit cell in size. Their binding
energy is on the order of 1 eV. The second type of excitons are called Mott-Wannier
excitons. These occur in materials with larger dielectric constants such as
semiconductors and other insulators. In this case the larger dielectric constant leads
to increased screening of the Coulombic attraction of the exciton by the material and
the exciton has a much larger radius. The electron and hole can be considered to be
orbiting their centre of mass with a radius of several unit cells. In addition the
binding energy of this type exciton is much smaller, on the order of 0.1 eV, than the
Frenkel type. This is the type of exciton formed in ZnO, which has an exciton
binding energy of ~ 60 meV and an electron-hole separation of ~ 2 nm.
64
The importance of excitonic effects is dependent on the size of the exciton
binding energy and the temperature at which experiments are performed. At
temperatures greater than the exciton binding energy (T > Ebinding/k) the electron and
hole are thermally dissociated and exciton effects are not observed, whereas at lower
temperatures (T < Ebinding/k) excitonic effects are clearly in evidence. For the case of
ZnO, where Ebinding = 60 meV, excitonic effects persist up to room temperature and
above (300 K 26 meV). Since the electron and hole motions are correlated in the
excitonic state and the electron and hole remain close to each other, the quantum
mechanical probability of radiative emission is increased for excitons compared to
uncorrelated carriers. The fact that ZnO is a direct band gap material with a strongly
bound excitonic state means that it is a very efficient light emitter at energies at and
below its band gap.
PL essentially relies on the recombinations of excitons in the material under
investigation and the corresponding emission of photons. We have discussed in this
section the case of a perfect and pure material, where the emitted photons have
energies characteristic of the material itself, specifically the material band gap and
FX binding energy. However, in less perfect or pure materials some or all of the
photons emitted will have energies characteristic of the defect or impurity at an
exciton has become bound, as described in the following section.
3.3 Defects in Semiconductors
In ZnO, the FX has an energy 60 meV below the band gap energy. However,
this work is more concerned with the case of bound excitons (BX). These are
excitons which have bound to defect or impurity sites in the crystal structure and are
no longer freely travelling through the crystal. PL is a powerful tool to use in the
analysis of defects and impurities as each BX has a unique localisation energy,
essentially its binding energy at that impurity, measured relative to the FX position,
which is specific to that defect or impurity. They therefore emit photons of a
specific energy upon recombination and can be unique fingerprints for identification
and study of the impurities/defects involved because the localisation energy depends
65
on the nature of the defect including its chemical identity, charge state, electronic
structure and the details of the nearby lattice. PL can be used to analyse the optical
spectrum of a material under optical excitation and can identify many peaks
associated with different defects or impurities.
Excitons may be bound at ionised donor sites (D+X) and neutral donor sites
(D0X) in ZnO.
3 These appear as peaks in the spectrum at energies just below the FX
position, and in that order for ZnO. The ordering for different defect charge states is
related to the electron and hole effective masses. These are characterised as shallow
defects and this region of the spectrum is called the band edge or near band edge
region. Donor-acceptor pair transitions (DAP) can also take place where an electron
at a donor site combines with a hole at an acceptor site.1 Two electron satellites
(TES) occur if a D0X recombines and leaves the donor atom in an excited 2s or 2p
state. Electrons in the conduction band can also combine with holes at acceptor
sites, as well as holes in the valance band combining with electrons at donor sites.
Figure 3.2 gives an illustration of FX, BX and possible transitions in the band gap.
The localisation energies of shallow BX in ZnO are in the range of ~ 3 to ~ 20 meV
and thus these excitons become thermally delocalised as temperature rises and are
not observed at room temperature, where only features due to FX are seen.
Excitons can also become bound at other defects and lead to emissions at
energies far from the band edge energy. This type of emission is referred to as deep
level emission (i.e. involving deep levels closer to the centre of the band gap) and the
concept of excitons is generally less useful in analysing such emission, which is
generally viewed in terms of electron and hole levels localised at the defect, although
the initial exciton process is often via the capture of an FX. Once again however the
emitted photons have a specific energy upon carrier recombination and are unique
fingerprints for identification and study of the impurities/defects involved because
the localisation energy depends on the nature of the defect including its chemical
identity, charge state, electronic structure and the details of the nearby lattice.
A key parameter of PL is that it allows the energy levels introduced in the
forbidden band gap by defects and impurities to be probed. Much more information
can be gathered using low temperature PL than making the measurement at room
temperature. This is for a number of reasons: (i) thermal effects can degrade PL
66
emission by delocalisation of excitons or carriers from defects and such effects are
reduced or eliminated at low temperatures, (ii) non-radiative processes become much
more likely at higher temperatures leading to much lower emission intensity from
BX at higher temperatures and (iii) because the spectral line broadening effects of
phonons are also reduced at low temperatures, and therefore working at low
temperatures leads to strong, narrow emission lines with specific and accurately
measureable localisation energies. In addition, low temperatures are therefore vital
to study such defects in detail.
Figure 3.2: (a) Schematic of a FX and D0X; (b) Representation of the main
transitions with PL: i) FX, ii) BX, iii) electron to acceptor, iv) donor to acceptor, v)
hole to donor.
In addition to radiative transitions such as FX and BX, when the entire energy
associated with the excited state energy is released as a photon, situations also occur
when exciton recombination takes place whereby some of the energy can be released
in non-radiative processes. This occurs when the energy of the exciton is converted
into vibrational energy of the lattice called phonons, with the emission coupled to the
crystal lattice vibrations, and a lower energy photon is emitted, called a Stokes
process (at higher temperatures there is also a possibility for phonon absorption
leading to higher energy photon emission – called an anti-Stokes process). This
type of coupling is strongest in the case of phonons known as longitudinal optical
(LO) phonons. A series of photons are emitted at reduced energies and appear in the
spectrum at specific intervals below the main emission which are whole numbers of
the phonon energy for the crystal structure. For ZnO, this energy is ~72 meV.5
67
These peaks are known as LO phonon replicas. Depending on the level of coupling,
described by the Huang-Rhys factor, the LO-phonon replica intensity can be quite
large compared to the ZPL intensity, particularly for deep defects near the centre of
the band gap.6 In direct band gap semiconductors, deep centres can produced bright,
broad band emissions for this reason. LO replicas are further discussed in section
3.5.
We finally note that levels near the centre of the band gap can be very efficient
non-radiative centres and therefore quench the optical emission by means of multi-
phonon de-excitation.7 This effect increases at higher temperatures, as noted above,
because BX are thermally delocalised to become FX and can diffuse and encounter
efficient non-radiative centres near the middle of the band gap.
3.4 Structure of the ZnO PL emission spectrum
Figure 3.3: Diagram showing the main types of bound exciton in the band edge
region of ZnO, reproduced from reference 1 (Note that I9 is a D0X, not an acceptor
bound exciton (A0X).
The BX recombinations discussed in the previous paragraph result in photon
emission at energies close to the band gap in the so called band edge and near band
edge region. Figure 3.3 shows the types of lines observed in this region in ZnO at
68
low temperatures and discussed in this and the previous sections. These lines appear
near the band gap in energy as they originate from shallow defect transitions
involving donor levels which are close to the conduction band or acceptor levels
close to the valence band. They are clearly visible in the band edge region of the PL
spectra of the ZnO samples in this work which are described in subsequent chapters.
The FX line in ZnO appears at ~3.377 eV, the band gap minus the exciton
binding energy. Two component states of the free A-exciton can be observed,
namely AL and the lower energy AT, corresponding to the longitudinal and
transverse polariton A-exciton energies respectively. The BX localisation energies
are measured from the AT energy. Moving towards lower energies, there is a series
of lines arising from D+X, and D
0X. This series of lines, labelled the I-lines, from I1
to I11, centred around 3.36 eV. These lines have been extensively studied and are
outlined in detail by Meyer et al. in reference 8 and chapter 7 of reference 1.
Briefly, I0 to I3 are due to D+X and I4 to I10 are due to D
0X lines. Not all the
lines have been conclusively identified. However of the main lines, I9 has been
shown to be caused by In impurities, as well as its ionised counterpart, I2.9,10
Likewise, Al impurities are the cause of the I6/I6a and I0 lines.8,11,12
I8 and I1 are
related to Ga,13
and I4 to H.14
Some of these lines are used to measure shifts in the
band edge (and therefore band gap energies) with changing isotopes in this work,
particularly the I9 peak which is the most pronounced in most of the samples used
here. The I-lines are very narrow as their full widths at half maximum (FWHM) are
reduced due to low thermal broadening at low temperatures and their lack of
translational motion at bound sites. Lattice strain can shift line positions1 and also
increase the FWHM (for inhomogeneous strains) so narrow lines indicate high
crystal structural quality. No A0X lines have been identified in ZnO so the next
features of the spectrum as we move to lower energies are TES of the I-lines. Since
this occurs when the recombination leaves the donor in an excited 2s or 2p state, the
features appear in the spectrum at an energy equal to the difference between this
excited state and the ground 1s state below the corresponding BX line. The other
main feature of this spectral region as we move to lower energies is the BX LO
replica region which shows features at intervals of 72 meV. The LO replicas of the
TES peaks can also be visible. Another series of lines appearing at lower energies
than the BX I-lines have also been observed in ZnO, labelled the Y-lines.15
These
69
are not as well characterised and understood, and they are not of interest in this
work.
Aside from the intense UV emission associated with these I-lines, the other
main feature of the PL spectrum of ZnO at low temperatures is the green band. This
is a broad emission centred in the green region of the visible spectrum around 2.4
eV. There are actually two green bands, known as unstructured and structured
(SGB). There is strong evidence that the unstructured green band is related to
VO16,17
, while the SGB has been shown to be caused by Cu impurities.18,19
This
emission is discussed in more detail in section 3.5. Spectra of both the band edge
and SGB regions are presented in section 4.6.
At room temperatures, the PL spectrum of ZnO has a few key differences.
The first is in the band edge region, where the BX I-lines are no longer visible (due
to thermal delocalisation of BX) and instead there tends to be a broader band edge
emission due to the FX recombinations, and their LO replicas, which are much more
numerous in this case. The exact form of this band and its peak position can vary
from sample to sample due to the varying exciton-phonon coupling in different
sample morphologies.20
The other main feature is a broad band emission in the
visible region. This is usually the unstructured green band, although yellow-orange
and red bands are also observed in ZnO at different temperatures, suspected to be
due to other metal impurities or native defects.21
The SGB is only seen at low
temperatures.
3.5 The Cu-related Defect at 2.86 eV
PL also provides the opportunity to examine the defects that cause deeper level
defects near the middle of the band gap. The Cu-related emission at 2.86 eV at low
temperatures in ZnO is one such emission centre and is of major interest in this
work. It produces a bright, structured, wide band emission peaking in the green
visible region, previously referred to as the SGB. It is made up of a ZPL at 2.86 eV
and a series of LO replicas at steps of ~72 meV showing strong electron-phonon
coupling. Although the emission spectrum of ZnO had been studied for some time,
70
the origins of this SGB were not certain when Dingle made one of the most
fundamental contributions to the discussion in 1969.18
He presented strong evidence
that this SGB resulted from a transition involving a single Cu impurity on a Zn site
(CuZn) in a Cu2+
charge state acting as a neutral deep acceptor. The ground state of
this transition is a tightly bound hole in a level slightly below the conduction band.
The excited state is at a level ~450 meV above the valence band.18,19
Dingle
observed that the ZPL was made up of two distinct lines, both of which behaved
identically under uniaxial stress and magnetic fields, and which had an intensity
radio which matched the isotopic distribution of the 63
Cu and 65
Cu isotopes in ZnO.
He also noted that one of the states involved had an anisotropic g-factor closely
matched to the g-factor of one of the states in a Cu2+
absorption transition reported
earlier. This emission also only appears when Cu impurity is present. Dingle
concluded that Cu was the cause of the SGB.
Despite Dingle’s work, and extensive studies concerning the green band,22–24
some disagreement persisted regarding its nature,21,25
and it was not until much more
recently that Cu was unambiguously confirmed by direct measurement as being the
cause of the SGB emission. This experiment was carried out in our group by Byrne
et al.19
ZnO single crystals were specifically doped using 63
Cu and 65
Cu isotopes.
The intensities of the respective parts of the ZPL doublet were observed to change
directly in proportion with the isotopic distribution of the Cu in a particular sample.
These data, in addition to Dingle’s data on the g-factor and stress measurements,
were consistent with the assignment of this defect as a single substitutional Cu2+
atom on a Zn site. However, Byrne et al. also noted that a complex of CuZn with
other defects along the c-axis would also reproduce this data, for example CuZn-VO
or CuZn-VZn complexes, and therefore the exact defect structure was not fully
determined. It has been suggested that Cu is present in ZnO as an unintentional
dopant in different charge states, lattice locations and in different complexes
including with native defects, the last of which is under investigation in this work,
involving the use of isotopically enriched ZnO nanorods to investigate whether this
defect is a complex involving any of these other native defects.26–28
It is also
reported that annealing at 900˚C may activate the Cu defect responsible for this
green emission by changing the charge state from Cu1+
to Cu2+
.29
Byrne et al.
observed this in ZnO nanorods and suggest this is because much of the Cu in the
71
nanorods following VPT growth is in the Cu+ charge state because of Zn vapour
reducing the Cu2+
to Cu+ during growth. Thus annealing at 900˚C for ten minutes
has been used in some samples in this work to increase the intensity of SGB in
samples were it was very weak relative to the band edge emission, or indeed to
‘activate’ it in cases where it was not observed in the samples as grown.
We note that the unstructured green band was also observed at low
temperatures in some samples, but it was clear that the two types of green band
emission have different origins. The unstructured band has been attributed to Zni
and VZn in the past;30,31
however there is more recent strong evidence that VO is the
cause of this emission.16,17
As mentioned in this chapter, the BX and deep defect emission processes can
couple to vibrations in the crystal lattice or phonons which results in LO replicas at
lower energies than the parent emission. The energy difference depends on the
vibration involved with 72 meV being characteristic of the ZnO lattice vibration.
This coupling can be described using a configuration coordinate diagram, as in
figure 3.4.6 For simplicity we consider only one mode of vibration and treat the
vibrating mode as a harmonic oscillator. Only one coordinate is needed to describe
the system in this case, labelled Q. This configuration coordinate Q may be thought
of in this case as representing the distance between the dopant atom and the nearest
adjacent atom, which may vary depending on the electronic and vibrational states the
atoms are in, resulting in different levels of overlap in their wavefunctions. Two
electronic energy levels are shown, a ground state and excited state, each with a
number of vibrational modes. Transitions may occur from the lowest phonon level
of the ground state, at Qog to lowest phonon level of the excited state, at Qoe, or to the
various phonon levels in the excited state. Instantaneous changes in the vibrational
modes with electronic transitions are possible because the electronic transition
occurs instantaneously compared to the slower motion of the atomic particles
(Frank-Cordon principle). In the figure for example the highest probability transition
is to the third phonon state, with the largest wavefunction overlap. The electron then
relaxes to the Qoe state by thermalisation and recombines back to the ground
electronic state. However, because of the difference in Q value, and therefore
wavefunction overlap, the recombination occurs with a higher phonon level in the
ground state (shown by solid lines in the figure). The emission is therefore at a
72
lower energy than the absorption. Indeed, when the oscillation frequencies of the
two states are equal, the entire emission spectrum occurs as a mirror image of the
absorption spectrum around the ZPL. The difference between the peaks of the
absorption and emission spectra is called Stokes shift.
Figure 3.4: Configurational coordinate diagram describing the ground and excited
states of an impurity and the absorption and emission spectra due electron-phonon
coupling.
There are a number of assumptions made in this model. The two electronic
states are assumed to modelled by a harmonic oscillator, it is assumed that at low
temperatures only the lowest phonon mode in the ground state is excited, and it is
assumed that both electronic states interact with the lattice in the same way. It is
clear therefore than this picture is simply an approximation and the actual behaviour
is much more complicated.
The wavefunction overlap between the zero phonon levels in the ground and
excited states, Qog and Qoe, is quite small and therefore less probable (shown as
dotted lines in the figure). However, the ZPL emission is made up of the sum of the
ZPLs from all the different vibrational modes which each contribute to the phonon
sideband in the PL spectrum. This occurs at the point marked E0. The level of
73
electron-phonon coupling is indicated by the Huang-Rhys parameter, S, which is
related to the oscillation frequency ω, the ion effective mass M and the displacement
of the upper and lower state parabola by the formula6
𝑆 =1
2
𝑀𝜔2
ℏ𝜔(𝑄0𝑒 − 𝑄0𝑔)
2 Eqn. 3.1
When S is zero or small, almost all the emitted energy emerges in the ZPL,
with little observed in phonon replica emission. When S is larger, replica peaks are
observed due to the coupling between defect electronic levels and lattice vibrations,
as in the band edge region of ZnO. When S is larger still, this sideband is broadened
resulting in a single, broad band, intense, structured emission, as in the Cu-related
SGB in ZnO, with characteristic LO replicas. Figure 3.5 gives an illustration of how
this band evolves as S increases.6 Note that this image refers to the absorption
spectrum, but serves as a good illustration of how the emission spectrum behaves,
but at energies below the ZPL rather than above in the absorption spectrum. The S
parameter for the Cu-related SGB in ZnO has been reported to be 6.5.29
Figure 3.5: Illustration of how the phonon sideband shape evolves with increasing S
parameter. Note this image refers to the absorption spectrum, but serves as a good
illustration of how the emission spectrum behaves but at energies below the ZPL.
Image reproduced from reference 6.
74
3.6 The Isotope Effect
Isotopic enrichment is a well-known and very useful technique in the analysis
of defects in semiconductor materials including ZnO, particularly in conjunction
with optical methods like PL, and has been studied extensively.32–34
The growth of
isotopically enriched ZnO nanorods in this work presented an opportunity to use
these samples in the study of defects and impurities, specifically the Cu-related
emission at 2.86 eV and its associated phonon sideband. The high optical quality of
the materials produced ultimately made this possible. A general description of the
origins of the isotope effect is given here and more discussion is presented alongside
the relevant results in later chapters. Briefly, the technique involves enriching the
crystal with either specific isotopes of an atom involved in, or suspected of
involvement in, a defect or enriching the surrounding native crystal lattice with
specific isotopes. The vibrational states in the crystal couple to electronic states,
including those at defect sites, to produce so-called vibronic levels via electron-
phonon coupling, as discussed above. Transitions between different vibronic levels,
either at a defect or the surrounding lattice produce the ZPL and phonon side-bands
involving multiple phonon replicas, as seen for the Cu-related SGB in ZnO.
Changes in the energies of exciton recombinations may arise due to changes in both
the isotopic make up of a defect or a surrounding lattice, both of which affect the
vibrational nature of the vibronic levels because of the difference in mass of different
isotopes.6,35
The isotope effect is fundamentally due to the dependence of the phonon
vibrational frequencies on atomic masses. If we consider a situation where a foreign
impurity atom vibrates in a particular mode with vibrational frequency ω, has a mass
m, and the vibration is characterised by a spring constant of k, then the frequency is
given by
𝜔 = √𝑘
𝑚 Eqn. 3.2
It can therefore be clearly seen that, assuming the spring constant (due to the
chemical nature of the binding and the interatomic forces produced by this binding)
does not change, a change in mass, to m+Δm for example, will result in a change in
75
the vibrational frequency. For an increase in mass for example, the phonon
frequency will reduce. When the upper and lower states have different spring
constant values this ultimately leads to a change in the energy of the transitions,
shown in the configurational coordinate diagram in figure 3.4, as the energy of the
coupled phonon modes changes. This is a simple case with many assumptions, as
described previously. A defect will couple to many vibrational modes in a real
crystal. The magnitude of the isotope shift is ultimately dependent on both the
difference in spring constant between the upper and lower levels and Δm. The
influence of the change in mass on the vibrational frequency produces the energy
change in the emission spectrum. The variation of phonon frequencies with average
mass and fluctuations about the average both contribute to the effect.22,35,36
As demonstrated by the work of Byrne et al.19
on the Cu-related 2.86 eV ZPL
mentioned above, isotopic substitution is a technique which can provide
confirmation of the involvement of a particular atom in a defect complex. In this
case the two isotopes of Cu appeared as two distinct lines in the ZPL. In addition to
isotopic shifts in this manner (where the impurity atom is present in multiple
isotopes, or enriched with a specific isotope) isotopes shifts have been observed in
ZnO when the isotopic content of the lattice is enriched. This can occur with
enrichment of either or both of the Zn or O atoms in the lattice. Shifts in the band
gap have been measured following enrichment of the lattice in this way.37,38
The
corresponding shifts in the FX and BX recombination energies have also been
measured using PL and reflectance spectroscopy.39,40
These reports have focused on
enrichment of single crystals of ZnO. In this work, the development of novel
methods of growing isotopically enriched ZnO in nanorod morphologies in a quick
and efficient manner presents an opportunity to use isotopically enriched
nanocrystals to carry out an optical study of defects in isotopically enriched crystals.
Specifically, a study of Cu-related defect emission with its ZPL at 2.86 eV was
carried out. The objective here was to determine if native defects were involved in
this defect, following the unambiguous confirmation of the Cu involvement but with
a possible complex with native defects not completely ruled out. By enriching the
nanorods as described in chapter 2, we can observe shifts in the energies of the
exciton recombinations in both the band edge and SGB ZPL. By changing the
masses of the Zn or O atoms in the lattice, the atomic surroundings of defects in both
76
regions would be altered in same way, and we can expect that the shifts observed
would be the same if no native defects complex with the CuZn. However, if a native
defect is involved in a complex with the CuZn defect, for example an additional
interstitial or vacancy, then its vibrational surroundings would be changed in an
additional way to the change induced by the lattice enrichment. Due to this, a
different energy shift may be observed for the defect SGB ZPL compared to that of
the band edge emissions.
3.7 Phonon modes
Phonon modes can occur in multiple types. The phonon modes which are
examined using Raman spectroscopy, and observed as replicas in PL, in this work
are described in this section. A more detailed discussion regarding phonon modes in
ZnO is presented in the literature 1,41
. As stated above, the LO replicas appearing in
the PL spectrum are the longitudinal optical type. Optical phonons are vibrational
modes which oscillate in an out-of-phase motion with adjacent atoms moving in
different directions. They can occur in either the longitudinal (LO) or transverse
(TO) direction relative to the direction of propagation. Optical phonons can interact
with incident light via infrared absorption or Raman scattering. The other type of
phonons is the acoustic modes labelled LA for the longitudinal mode and TA for the
optical mode. Acoustic phonon modes have a coherent oscillation with adjacent
atoms moving in the same direction.3
The primary optical phonon modes of interest in this work are shown in figure
3.6. They are labelled as the E2low
and E2high
modes. The E2low
mode is dominated
by the Zn atom motion, and the E2high
mode is dominated by the O atom motion. The
ratio of their displacement eigenvector lengths, eZn/eO, have been reported as -2.4
and 0.417 respectively.42
In each case, both the Zn and O atoms oscillate in opposite
directions perpendicular to the c-axis. Both modes are non-polar as the
compensating displacements of each sub-lattice (Zn or O) leads no overall dipole
moment. That is, each Zn or O atom moves in the opposite direction to the
neighbouring one at any given time, producing no net electric field, and therefore no
77
significant coupling to excitons. Since there is no difference between the
longitudinal and transverse modes in these cases, they are not referred to as LO or
TO. For the backscattered Raman geometry used in this work as described in section
2.4.4, the E2low
mode appears in the Raman spectrum at ~100 cm-1
and the E2high
mode appears at ~439 cm-1
. Raman measurements of phonons in isotopically
enriched single crystal ZnO have been carried out previously, particularly on the
E2high
mode, clearly showing the changes in phonon energies and line widths with
changing Zn or O isotopes.43–45
Similar shifts are also seen in similar materials such
as GaN.46
It is therefore useful to compare Raman measurements made with
enriched ZnO nanorods with these previous data, as they can provide a good way
confirming the isotopic enrichment of the samples, along with comparisons of the
BX line energies using PL.
The LO phonon replicas observed in both the band edge and SGB emission in
ZnO are, however, not the E2high
or E2low
modes. These replicas are due to the E1
(LO) and A1 (LO) phonon modes, which have energies of 590 cm-1
and 574 cm-1
respectively (they each have TO counterparts at lower energies).1 These two modes
are polar as the Zn and O sub-lattices essentially move in opposite directions at any
given time. This leads to significant electric fields which strongly couple to excitons
in the crystal and cause the LO replicas in the PL spectrum. Note that the energies of
these polar modes are 73.2 meV for E1 (LO) and 71.2 meV for A1 (LO) which match
the PL LO replica spacing closely. Since these modes are much weaker in intensity
and do not appear in the same Raman geometry as the E2 modes, and since previous
studies on the E2high
mode in isotopically enriched ZnO are available, the E1 and A1
modes are not considered further. Both modes are also shown in figure 3.6.
We also note that the measurement of Raman mode shifts provides a very
reliable measurement of sample isotopic enrichment, since the Raman phonon mode
energies are much less sensitive to any lattice strain than are PL peaks. For example,
Wagner states figures for energy shifts of 3.37 meV/GPa for the I9 line and just 0.35
meV/GPa (< 3 cm-1
/GPa) for the E2high
phonon energy.41
78
Figure 3.6: E2low
, E2high
, E1 and A1 phonon modes in wurtzite ZnO. Image adapted
from reference 1.
3.8 Conclusions
The background principles of low temperature PL, emission from
semiconductors and defects in semiconductors, the isotope effect and the phonon
modes of interest for Raman studies have been introduced in this chapter. The main
emissions typically observed in the low temperature PL spectrum of ZnO were
described, particularly the band edge region and the Cu-related SGB, in addition to
the processes underlying such emissions. This background information will be used
in subsequent chapters alongside the presentation of results from studies on Zn- and
O- enriched ZnO nanorods, and a detailed optical study of the 2.86 eV Cu-related
defects using those materials.
3.9 References
1 C.F. Klingshirn, B.K. Meyer, A. Waag, A. Hoffmann, and J. Geurts, ZnO: From
Fundamental Properties Towards Novel Applications (Springer-Verlag, Berlin,
Heidelberg, 2010).
2 A. Mang, K. Reimann, and S. Rübenacke, Solid State Commun. 94, 251 (1995).
3 C. Klingshirn, Semiconductor Optics, 2nd ed. (Springer-Verlag, Berlin, Heidelberg,
2005).
79
4 P.Y. Yu and M. Cardona, Fundamentals of Semiconductors, 3rd ed. (Springer-
Verlag, Berlin, Heidelberg, 2001).
5 L. Wang and N.C. Giles, J. Appl. Phys. 94, 973 (2003).
6 B. Henderson and G.F. Imbusch, Optical Spectroscopy of Inorganic Solids
(Clarendon Press, Oxford, 1989).
7 E. Rashba and M. Sturge, Excitons (North-Holland Publishing Company,
Amsterdam, 1982).
8 B.K. Meyer, H. Alves, D.M. Hofmann, W. Kriegseis, D. Forster, F. Bertram, J.
Christen, A. Hoffmann, M. Straßburg, M. Dworzak, U. Haboeck, and A. V. Rodina,
Phys. Status Solidi 241, 231 (2004).
9 S. Muller, D. Stichtenoth, M. Uhrmacher, H. Hofsass, C. Ronning, and J. Roder,
Appl. Phys. Lett. 90, 012107 (2007).
10 J. Cullen, D. Byrne, K. Johnston, E. McGlynn, and M.O. Henry, Appl. Phys. Lett.
102, 192110 (2013).
11 M. Schilling, R. Helbig, and G. Pensl, J. Lumin. 33, 201 (1985).
12 B.K. Meyer, J. Sann, S. Lautenschläger, M.R. Wagner, and A. Hoffmann, Phys.
Rev. B 76, 184120 (2007).
13 K. Johnston, M.O. Henry, D. McCabe, E. McGlynn, M. Dietrich, E. Alves, and M.
Xia, Phys. Rev. B 73, 165212 (2006).
14 E. Lavrov, F. Herklotz, and J. Weber, Phys. Rev. B 79, 165210 (2009).
15 M.R. Wagner, G. Callsen, J.S. Reparaz, J.-H. Schulze, R. Kirste, M. Cobet, I. a.
Ostapenko, S. Rodt, C. Nenstiel, M. Kaiser, A. Hoffmann, a. V. Rodina, M.R.
Phillips, S. Lautenschläger, S. Eisermann, and B.K. Meyer, Phys. Rev. B 84, 035313
(2011).
16 F.H. Leiter, H.R. Alves, A. Hofstaetter, D.M. Hofmann, and B.K. Meyer, Phys.
Status Solidi 226, R4 (2001).
17 F. Leiter, H. Alves, D. Pfisterer, N.G. Romanov, D.M. Hofmann, and B.K. Meyer,
80
Phys. B Condens. Matter 340-342, 201 (2003).
18 R. Dingle, Phys. Rev. Lett. 23, 579 (1969).
19 D. Byrne, F. Herklotz, M.O. Henry, and E. McGlynn, J. Physics. Condens. Matter
24, 215802 (2012).
20 T. Voss, C. Bekeny, L. Wischmeier, H. Gafsi, S. Borner, W. Schade, a. C. Mofor,
A. Bakin, and A. Waag, Appl. Phys. Lett. 89, 182107 (2006).
21 H. Morkoç and U. Özgür, Zinc Oxide Fundamentals, Materials and Device
Technology (Wiley-VCH, Weinheim, 2009).
22 V. Heine and C. Henry, Phys. Rev. B 11, 3795 (1975).
23 J.A. Van Vechten, Phys. Rev. B 13, 946 (1976).
24 P. Dahan, V. Fleurov, P. Thurian, R. Heitz, A. Hoffmann, and I. Broser, J. Phys.
Condens. Matter 10, 2007 (1998).
25 U. Ozgur, Y.I. Alivov, C. Liu, a. Teke, M. a. Reshchikov, S. Dogan, V. Avrutin,
S.-J. Cho, and H. Morkoc, J. Appl. Phys. 98, 041301 (2005).
26 F. Herklotz, E.V. Lavrov, and J. Weber, Phys. B Condens. Matter 404, 4807
(2009).
27 H. Qiu, F. Gallino, C. Di Valentin, and Y. Wang, Phys. Rev. Lett. 106, 066401
(2011).
28 U. Wahl, E. Rita, J.G. Correia, T. Agne, E. Alves, and J.C. Soares, Superlattices
Microstruct. 39, 229 (2006).
29 N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, and G. Cantwell,
Appl. Phys. Lett. 81, 622 (2002).
30 N.O. Korsunska, L.V. Borkovska, B.M. Bulakh, L.Y. Khomenkova, V.I.
Kushnirenko, and I.V. Markevich, J. Lumin. 102-103, 733 (2003).
31 X. Yang, G. Du, X. Wang, J. Wang, B. Liu, Y. Zhang, D. Liu, D. Liu, H.C. Ong,
and S. Yang, J. Cryst. Growth 252, 275 (2003).
81
32 M. Cardona, Phys. Status Solidi 220, 5 (2000).
33 V.G. Plekhanov, Applications of the Isotope Effect in Solids (Springer-Verlag,
Berlin, Heidelberg, 2004).
34 V. G. Plekhanov, Isotopes in Condensed Matter (Springer Berlin Heidelberg,
Berlin, Heidelberg, 2013).
35 E. McGlynn, A Photoluminesence Study of Cadmium and Aluminium-Related
Defects in Silicon, Dublin CIty University, 1996.
36 V.G. Plekhanov, J. Nucl. Sci. Technol. 43, 375 (2006).
37 F.I. Kreingol’d and B.S. Kulinkin, Fiz. Tvend. Tela 28, 3164 (1986).
38 F.I. Kreingol’d, Fiz. Tvend. Tela 20, 3138 (1978).
39 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
Cardona, Solid State Commun. 128, 35 (2003).
40 S. Tsoi, X. Lu, A.K. Ramdas, H. Alawadhi, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 74, 165203 (2006).
41 M.R. Wagner, Fundamental Properties of Excitons and Phonons in ZnO: A
Spectroscopic Study of the Dynamics, Polarity, and Effects of External Fields,
Technical University of Berlin, 2010.
42 J. Serrano, a. Romero, F. Manjón, R. Lauck, M. Cardona, and a. Rubio, Phys.
Rev. B 69, 094306 (2004).
43 J. Serrano, F. Manjón, A. Romero, F. Widulle, R. Lauck, and M. Cardona, Phys.
Rev. Lett. 90, 055510 (2003).
44 J. Serrano, F. Widulle, A.H. Romero, A. Rubio, R. Lauck, and M. Cardona, Phys.
Status Solidi 235, 260 (2003).
45 M.A. Gluba, N.H. Nickel, and N. Karpensky, Phys. Rev. B 88, 245201 (2013).
46 J.M. Zhang, T. Ruf, M. Cardona, O. Ambacher, M. Stutzmann, J.-M. Wagner, and
F. Bechstedt, Phys. Rev. B 56, 14399 (1997).
82
Chapter 4: Zn Isotope-enriched ZnO
Nanorods
4.1 Introduction
This chapter presents the results of the growth and characterisation of ZnO
nanorods enriched with Zn isotopes as described in section 2.2. The results of
characterisation of the morphology, crystal quality, isotopic enrichment and optical
quality of the samples using SEM, XRD, SIMS, Raman spectroscopy and low
temperature PL are presented as well as some reflectance data. The eight Zn
isotopically enriched samples of ZnO nanorods produced for this study are shown in
table 4.1. Figure 4.1 displays a graphical representation of the Zn isotopic content of
this set of samples. The mixed samples’ labels refer to the nominal ratios of the
isotopes by mass as the source powder was weighed during growth.
Figure 4.1: Graphical representation of the Zn isotopic content of the eight
Zn-enriched samples.
83
Label Sample description
64ZnO
64ZnO sample
66ZnO
66ZnO sample
68ZnO
68ZnO sample
64/66ZnO
1/2
64ZnO and
1/2
66ZnO
66/68ZnO
1/2
66ZnO and
1/2
68ZnO
64/68ZnO
1/2
64ZnO and
1/2
68ZnO
64/66/68ZnO
1/3
64ZnO,
1/3
66ZnO and
1/3
68ZnO
natZnO Natural Zn isotopic distribution, avg. =
65.4ZnO
Table 4.1: Set of Zn-enriched ZnO nanorods.
Figure 4.2: Typical buffer layer of CBD deposited nanorods.
As described in section 2.2, these samples were grown using a modified VS
CTR-VPT technique on pre-deposited unenriched nat
ZnO CBD buffer layers, based
on the three-step process previously developed in our group.1 All of the samples in
this set were grown using the NaOH-based chemical bath as proposed by Peterson et
al.2 The CBD buffer layer consists of a uniform film of c-axis aligned ZnO
nanorods of a few hundred nanometres in length on top of the seed layer as proposed
by Greene et al.3 Figure 4.2 shows a typical NaOH-method CBD layer. The CBD
layer provides energetically favourable sites for subsequent nucleation of the VPT-
84
grown c-axis aligned and isotopically enriched nanorods in the next step as ZnO will
not nucleate on unseeded Si substrates. The chemical origins of the seed layer have
been described by Byrne et al. as well as the three-step growth process in general.1
This CBD bath was used most often in this work because of the lack of precipitates
produced by the process.
The VPT step following CBD deposition involves the use of carbon to reduce
the ZnO to metal vapour and carbon monoxide according to equation 4.1.
𝑍𝑛𝑂(𝑠) + 𝐶(𝑠) → 𝑍𝑛(𝑔) + 𝐶𝑂(𝑔) Eqn. 4.1
The reaction then proceeds as a VS process when the Zn vapour condenses and is re-
oxidised at the energetically favourable sites provided by the CBD buffer layer. ZnO
grown on non-epitaxially matched substrates is typically not aligned, however since
the CBD layer is c-axis aligned, the VPT grown nanorods are also aligned in such a
way. The reaction shown in equation 4.1 dominates at temperatures above ~700°C.
Below these temperatures a similar reaction producing mainly CO2 dominates. The
CTR produces Zn vapour at temperatures much lower than those at which significant
direct sublimation of ZnO occurs, as discussed in chapter 1. A key aspect to note
concerning the CTR reaction in this VPT step is that the oxygen taking part in re-
oxidation of the Zn vapour at the CBD nucleation sites is residual O2 in the furnace
quartz tube from atmosphere. Most of this O2 is removed by the 5-10 minutes Ar
flush at the start of the VPT step but a sufficient amount remains for this process. In
chapter 5, the growth of O-isotopically enriched ZnO nanorods requires the
development of a new growth method involving the removal of this residual O2.
Note also that, looking at this reaction, replacement of the ZnO source powder with
Zn-isotopically enriched source powders should result in correspondingly enriched
nanorods.
The first tests required in modifying VPT growth step for such Zn-enriched
nanorods was to determine if the VPT process in our system could be carried out
using much smaller amounts of source powders, i.e. 10 mg each of ZnO and
graphite, as opposed to the usual 60 mg of each used in this process in our group
previously. This was a necessity simply because limited amount of the 64
ZnO,
66ZnO and
68ZnO powders were available, although it also has the obvious advantage
of producing the desired samples at lower cost. Figure 4.3 shows nat
ZnO nanorods
85
grown using (a) 60 mg and (b) 10 mg of each powder. Growth was successfully
carried out using the smaller amount. This was perhaps not unexpected as
significant amounts of the source powders are typically left in the boat with a ‘crust’
present at the end of the process. This is thought to be due to re-deposition of ZnO
on the powder mixture forming a layer with no carbon present and limiting the
reaction. Following this, a number of tests, using the reduced 10 mg of each powder,
were carried out at slightly different growth temperature ranging from 850-950°C.
Nanorods were observed in these cases, and a temperature of 925°C was
subsequently chosen at which to carry out the isotopically enriched growths.
Finally, it is worth noting that there is a temperature overshoot in the furnace of
about 100°C above the set point when heating with the temperature then falling to
about 50°C above the set point over about 10-15 minutes.
Figure 4.3: VPT-grown ZnO nanorods with (a) 60 mg and (b) 10 mg of each source
powder at 900°C.
4.2 Morphology - SEM
The growth procedure outlined in chapter 2 and section 4.1 produced ZnO
nanorods on the Si substrates. The morphology was examined using SEM. The Si
substrates and buffer layers have been covered with a dense array of nanorods in all
of the samples in the set. The vertical nanorods are well aligned along the c-axis due
to the preferential growth in this direction encouraged by the seed and buffer layer
textures. Coverage of nanorods on the substrates is generally excellent, with good
86
growth occurring over the whole substrate, except for the edges in some cases where
the sample can be slightly over the edges of the alumina boat and therefore blocked
from having nanorods deposited as these areas were not exposed to Zn vapour. The
following figures provide examples of the typical morphology seen in the samples.
They are imaged from above, at 30˚ to the vertical and at 90˚ to the vertical
producing cross-sectional images.
Figure 4.4: SEM images showing typical morphology of the ZnO nanorods: (a) plan
view, (b) 30˚ to the vertical and (c) 90˚ to the vertical. Images are from the 64
ZnO
sample.
87
As shown in figure 4.4(a), the top of the nanorods are clearly observed.
Viewing from an angle of 30˚ to the vertical, figure 4.4(b) displays the nanorods
clearly along a freshly cleaved edge. The nanorods are shown in the most detail in
figure 4.4(c). The cross section is shown, again along a freshly cleaved edge through
the centre of the sample. The nanorods are approximately 1-2 µm in height,
although the variation in a single sample is much smaller than this range. Coverage
of the substrate is widespread and dense, and the rods are very well aligned
vertically, along the c-axis. These images are from the 64
ZnO sample but similar
morphology is observed in all other samples. Previous studies carried out in our
group investigated the relationship between the nanorod height and diameter.4
Nanorods of length 1-2 µm typically have diameters of ~100 nm, with shorter
nanorods having larger diameters than longer ones. Based on simple observation of
the SEM images in this section, the samples in this work appear to fit these typical
dimensions.
Figure 4.5: (a) 90˚ angle of 66/68
ZnO nanorods with CBD buffer layer visible; (b)
64/66ZnO nanorods with CBD buffer layer visible at 90˚; (c) CBD buffer layer at 90˚
in 64/66
ZnO; and (d) 30˚ view of longer nanorods on part of 66
ZnO which have lost
their alignment and become entangled;
88
Figure 4.5 shows examples of nanorods with isotopic distributions labelled
according to table 4.1 as 66
ZnO, 64/66
ZnO and 66/68
ZnO. The CBD buffer layer is
clearly visible in figures 4.5(a) and 4.5(b) with VPT grown nanorods growing from
the nucleation sites provided by the buffer layer nanorods. The vertical c-axis
alignment is clearly observed. Figure 4.5(c) shows the CBD buffer layer on its own.
This image was taken on a part of sample which was over the side of the alumina
boat during growth, thereby preventing VPT phase growth from occurring at this
location. Figure 4.5(d) is taken from a part of the 66
ZnO sample. In this case the
nanorods are much longer than usual, on the order of a few tens of μm, and they
appear to have lost their alignment in the upper parts of the nanorods after falling
over on each other and have become entangled. At stages of this work, it has been
noted that sometimes this type of morphology is observed. It is suggested that this
could be related to the Ar flush time at the start of the VPT step, as it was noticed
that this occurred a number of times when a slightly shorter flush time was used,
specifically during the growth of O-enriched nanorods in DCU discussed further in
chapter 5. The shorter Ar flush leads to a larger residual O2 concentration and
therefore a slightly longer nanorod morphology.
Figure 4.6: 64/66/68
ZnO samples at 30˚ (a and c), 90˚ (b and d). The first sample (a
and b) shows poor nanorod morphology. The second sample (c and d) has improved
nanorod morphology.
89
The first 64/66/68
ZnO sample displayed poor nanorod morphology as shown in
figure 4.6 (a and b). There is very little nanorod growth during the VPT phase and
any rods are mostly clumped together in groups, producing a mostly random pattern
with no particular structure. It appears that in this case a relatively thin layer of
material was deposited on the CBD layer during VPT growth and this has poor
nanorod morphology. Visually, the morphology is more similar to the CBD layer
itself rather than aligned nanorods usually grown during the VPT step. The growth
of this sample was therefore repeated. In the second 64/66/68
ZnO sample, the nanorod
morphology is notably improved as shown in figure 4.6 (c and d). In this sample the
nanorods are clearly visible although there is still some clumping of rods together.
The VPT nanorods are of a slightly shorter length than the other samples also, as
seen in figure 4.6(d).
As the SEM data above has illustrated, the growth of isotopically enriched
ZnO nanorods utilising the modified three-step process previously used has been
very successful. This has been achieved by simply substituting the VPT source
powder for isotopically enriched powder and implementing the practical step of
reducing the amount of powder used. This easy, reliable and inexpensive method
has produced c-axis aligned nanorods of a very high quality with dense coverage
over a wide substrate area in each sample. This makes them ideal for use in optical
studies such as PL as described later. Some variations in morphology in terms of
length and width can be seen in the figures above and occur from time to time, and
indeed occasionally the morphology does not resemble nanorods as shown in figure
4.6(a). The growth process is sensitive to a number of parameters, including
temperature, distance of the sample from the powder and the source powder mixing
and particulate size. It is believed that the process of grinding the powders together
is vital and that variations in the amount of contact between the powders, as well as
temperature, may cause the growth reaction to occur at slightly different rates. This
may be responsible for the variations in morphology observed, for example in figure
4.6. Despite this, the batch-to-batch reliability of the growth method is very good,
and most of the samples produce nanorods as shown in figure 4.5 over most of their
surface area.
90
4.3 Alignment and Crystal Quality - XRD
XRD measurements were carried out on all samples in order to confirm the
presence of ZnO following VPT deposition and to study the crystal quality and
alignment. The x-ray source was a Cu Kα line with an effective mean wavelength of
0.15418 nm.
Figure 4.7: 2θ-ω spectra of several Zn-enriched ZnO samples dominated by the Si
substrate peak at 69.1˚ and the ZnO peak at 34.4˚. A number of other features are
observed as described in the text.
Figure 4.7 shows the XRD 2θ-ω diffractograms for the nat
ZnO, 64
ZnO, 66
ZnO
and 68
ZnO samples. These curves are typical of those recorded for the other samples
in the set. The 2θ-ω diffractograms are dominated by the 69.1˚ Si peak from the
(004) Si planes of the substrate and the 34.4˚ ZnO peak from the (0002) planes in the
deposited ZnO nanorods.5 The second order (0004) ZnO planes are also present.
6
The kinematically forbidden Si (002) is also present in some samples due to double
diffraction effects.7 A number of smaller peaks are observed and are attributed to
91
the plastic backed adhesive tape used to mount the samples on the stage. These are
labelled as T. A number of features are also visible due to Kβ radiation at ~62˚ from
the x-ray tube and tungsten Lα radiation at ~66˚ from contamination of the x-ray tube
Cu target by the electron gun tungsten filament.6 These features are marked as X.
The ZnO (0002) is far more intense than the other ZnO reflections, indicating a high
degree of alignment of nanostructures with their c-axes normal to the substrate
surface. This is consistent with the SEM images of the samples presented
previously. XRD therefore confirms the presence of wurtzite ZnO on the substrate
as a result of the growth reactions.
Table 4.2: ZnO (0002) peaks, FWHM and crystallite size and ZnO c lattice
constants, and Si (004) peaks and lattice constants in Zn-enriched nanorods.
Sample
ZnO
(0002)
peak
2θ (°)
ZnO
(0002)
FWHM
2θ (°)
Crystallite
size (nm)
ZnO c
lattice
constant
(nm)
Si
(004)
peak
(°)
Si
lattice
constant
(nm)
64ZnO 34.43 0.211 44 0.521 69.12 0.544
66ZnO 34.43 0.209 45 0.521 69.10 0.544
68ZnO 34.45 0.224 41 0.521 69.12 0.544
64/66ZnO 34.47 0.219 43 0.520 69.14 0.543
66/68ZnO 34.45 0.214 44 0.521 69.13 0.544
64/68ZnO 34.47 0.219 43 0.520 69.17 0.543
64/66/68ZnO
34.49 0.221 42 0.520 69.13 0.544
natZnO 34.43 0.203 47 0.521 69.12 0.544
92
Table 4.2 gives the ZnO (0002) and Si (004) 2θ peak positions, the FWHM
of the ZnO (0002) peak and calculated lattice constants for each sample as calculated
using equation 2.1. The average value for the c lattice constant in ZnO was 0.521
nm which agrees with the previously measured value of 0.521 nm stated in chapter
1.8 Likewise, the average value for the Si lattice constant was found to be 0.544 nm.
This is again in good agreement with the reported value of 0.543 nm for the lattice
constant in pure Si.9 An error of ±0.025° in the 2θ angles leads to an error in these
lattice constants of ±0.004 nm for the ZnO c lattice constant and ±0.002 nm for the
Si lattice constant. The narrow FWHM (~0.20-0.22° for all eight samples) of the
34.4˚ ZnO peaks indicates the high crystal quality. This is consistent with the
findings from SEM characterisation showing well aligned high quality nanorods
described previously. We can calculate the average out-of-plane nanocrystallite size
using the Scherrer relation10
:
𝐷 =𝐾𝜆
𝛽ℎ𝑘𝑙 cos 𝜗 Eqn. 4.2
where D is the crystallite size, K is a shape factor equal to 0.9, λ is the x-ray
wavelength, θ is half the 2θ angle and βhkl is the peak width when the instrumental
peak broadening of ~0.10˚ is taken into account by the formula
𝛽ℎ𝑘𝑙 = √(𝛽ℎ𝑘𝑙)𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑2 − (𝛽ℎ𝑘𝑙)𝑖𝑛𝑠𝑡𝑟𝑢𝑚𝑒𝑛𝑡𝑎𝑙
2 Eqn. 4.3
These widths measured imply crystallite coherence lengths of ~41-47 nm from the
Scherrer relation with an error of ±6 nm. The figures for each sample are given in
table 4.2.
The rocking curves of the 34.4˚ (0002) ZnO peak from the nat
ZnO, 64
ZnO,
66ZnO and
68ZnO samples are shown in figure 4.8. The peak positions and FWHM
of the rocking curves of all samples are shown in table 4.3. The high quality of the
crystal structure and c-axis alignment shown in SEM images and 2θ-ω scans above
is confirmed by the narrow FWHM of the rocking curves. The rocking curves peak
at ~17˚ and have FWHM of 2.26-3.25˚ for all eight samples. The peak positions
vary slightly due to small variations in sample tilt on the stage, due to the mounting
process.5 Again, in this case, the curves shown in figure 4.8 are typical of the other
samples in the set. Finally, it is interesting to comment on the 2θ-ω diffractograms
93
of the CBD buffer and seed layers as shown in figure 4.9. The buffer layer displays
the ZnO (0002) and (0004) peaks, but only the Si substrate peaks are visible for the
seed layer. The inset shows rocking curve of ZnO (0002) peak in the CBD nanorods
which had a FWHM of 5.68°, around twice that of the VPT nanorods, indicating that
the buffer layer is not as well textured as the VPT nanorods grown subsequently.
Figure 4.8: XRD rocking curves of the 34.4˚ (0002) ZnO peak from the nat
ZnO,
64ZnO,
66ZnO and
68ZnO samples.
Table 4.3: Peaks and FWHM of the rocking curves of the 34.4˚ ZnO (0002) peak in
all samples.
Sample Rocking curve peak
(°) FWHM (°)
64ZnO 16.91 2.4
66ZnO 16.8 2.26
68ZnO 16.84 2.92
64/66ZnO 17.15 2.67
66/68ZnO 16.95 2.88
64/68ZnO 17.49 2.56
64/66/68ZnO
17.04 2.57
natZnO 17.25 3.25
94
Figure 4.9: 2θ-ω diffractograms of CBD buffer layer nanorods and ZnO seed layer
only. (Inset shows rocking curve of ZnO (0002) peak in CBD nanorods.)
The SEM and XRD data have illustrated that the growth of ZnO nanorods
utilising this modified three-step process is very successful in terms of sample
crystallinity and morphology. This has been achieved by substituting the normal,
natural isotope content ZnO source powder used in VPT with isotopically enriched
powder and implementing the practical step of reducing the amount of powder used.
This easy, reliable and inexpensive method has produced well-aligned nanorods of a
very high quality with dense coverage over a wide substrate area. This makes them
ideal for use in further optical studies such as Raman and PL.
4.4 Isotopic Enrichment - SIMS
SIMS measurements were carried out on the set of eight samples in order to
obtain a direct measurement of the isotopic enrichment on two systems as described
in section 2.4.3. Figures 4.10 and 4.11 display the spectra from both systems which
are referred to as ‘SIMS’ and ‘miniSIMS’ in order to distinguish them.
Measurements on the ‘SIMS’ system were carried out by Mr. Conor Byrne in DCU.
SIMS in general provides key evidence as to whether isotopic enrichment has been
successful during the growth procedure by measuring the masses of the Zn atoms in
the samples. The region of interest here is around 64 to 68 amu, as we are measuring
95
the Zn+ ions ejected from the samples under the respective incident ion beams. Each
figure shows (a) nat
ZnO, (b) 64
ZnO, (c) 66
ZnO, (d) 68
ZnO, (e) 64/66
ZnO, (f) 66/68
ZnO,
(g) 64/68
ZnO and (h) 64/66/68
ZnO. Red, green and blue lines show the positions of 64,
66 and 68 amu respectively. The samples have been shifted slightly on the x-axis to
correct for a slight calibration error misalignment.
The first thing to note is that, in the nat
ZnO samples in both cases, the three
main isotopes, 64
Zn, 66
Zn and 68
Zn, follow the pattern that would be expected due to
the isotopic composition of nat
ZnO (48.6% 64
Zn, 27.9% 66
Zn, 18.8% 68
Zn). Looking
at the spectra for isotopically pure samples, 64
ZnO, 66
ZnO and 68
ZnO, it is clear that
the expected Zn isotope is dominant in each respective sample, thereby showing that
the samples have been enriched to a very high level. For each enriched isotope,
there are much smaller peaks at the masses of the other two isotopes. For example,
in figure 4.9(b) for 64
ZnO, small peaks are observed at 66 and 68 amu. These small
peaks are attributed to some ions of these isotopes arising from the CBD buffer
layer, which is not isotopically enriched, and to other ions from any surface
contaminants at these masses. The samples were tilted during these measurements,
as described in section 2.4.3, in order to reduce effects from the underlying substrate,
but even with this precaution the VPT nanorod coverage is not complete and a small
fraction of the buffer layer is exposed to the ion beam. Despite this, it is clear from
the these data that the nanorod samples have been successfully isotopically enriched
to a very high level using this growth method.
In the mixed isotope samples (64/66
ZnO, 66/68
ZnO, 64/68
ZnO, 64/66/68
ZnO), the
respective isotopes in each mixture are clearly dominant, with the isotope not
included in each growth not detected in significant amounts. This confirms the
successful enrichment of the mixed isotope samples, although the isotopes present
are not always measured to be present in exactly the desired proportions (1/2:
1/2 or
1/3:
1/3:
1/3, particularly in figure 4.10 (h)). This could be a result of some ions from
the underlying buffer layer being detected again, although it could also indicate that
the mixed samples did not grow in exactly these proportions, as the
disproportionality seems to be greater than the signal observed for isotopes not
present in the VPT growth for 64
ZnO, 66
ZnO and 68
ZnO. This is discussed further in
chapter 6.
96
Figure 4.10: SIMS spectra from each sample in the Zn-enriched set.
97
Figure 4.11: miniSIMS spectra from each sample in the Zn-enriched set.
98
It can be noted than although the samples were tilted as mentioned to try to
avoid signal from the buffer layer, tests performed with ion beam normal to the
surfaces did not produce notably different results.
Finally it must be noted that, in the miniSIMS spectra in figure 4.11, the other
main features are additional peaks observed to the right of each isotope peak, at one
amu greater. These are attributed to ZnH+ ions
11, and they appear to mostly follow
the behaviour of their ‘parent’ Zn+ peaks. Some small deviations from this pattern
could be the result of the different amounts of these ions generated in the system or
to some other species appearing at these masses. Indeed, the miniSIMS systems also
displays some other peaks at 63 amu, 71 amu and 73 amu which could be the result
of such species either from the sample or remaining in the chamber as the vacuum
level here was lower than the SIMS system. We note also that the Ga ion beam in
the miniSIMS systems produces isotopes of 69
Ga and 71
Ga, which could also
contribute to these peaks. Overall, the SIMS results show conclusively that the
novel VPT method developed here has successfully produced a set of isotopically
pure ZnO nanorods enriched to a very high level. Data from both the SIMS and
miniSIMS systems are consistent and this is consistent with the Raman results and
PL spectra below.
4.5 Phonon Frequencies - Raman
Data from Raman spectroscopy measurements performed on 64
ZnO, 66
ZnO and
68ZnO samples at room temperature are presented in figure 4.12. These
measurements were performed by Dr. Joseph Cullen, in Linköping University in
Sweden. Raman measurements were carried out in order to measure the Raman
frequencies in samples with the different isotopic purities and compare them to
previous measurements on enriched single crystals in the literature. In the inset of
figure 4.12, the Raman signal from the underlying Si substrate at around 520 cm-1
is
shown for reference.
For all three samples, the Raman spectra contain first-order E2low
and E2high
phonon modes that are typical for crystallite wurtzite ZnO as previously described in
99
chapter 3. The two Raman modes gradually shift to lower frequencies with
increasing Zn isotope mass from 64
ZnO to 68
ZnO, by 1.81cm-1
for E2low
and 2.17cm-1
for E2high
. These figures are consistent with previous findings of shifts of 1.59 cm-1
for E2high
at low temperatures (~6 K).12
The frequencies of the E2low
and E2high
phonon modes for 64
ZnO were 100.60 cm-1
and 439.03 cm-1
respectively. The
frequencies for each sample and their FWHM are summarized in table 4.4.
Shifting phonon frequencies for O and Zn isotopic substitution in ZnO have
been observed previously in a number of papers by Serrano et al.12,13
and are
observed here for the E2high
and E2low
phonon modes. Since only the natural
abundance of O (>99% 16
O) was used during nanorod growth in this work the
changing Zn mass is solely responsible for the shifting phonon frequencies. For the
E2high
phonon mode the observed frequency change for different Zn masses (around
2 cm-1
) is less pronounced than that previously reported for different O masses
(around 20 cm-1
) due to the dominating O eigenvector.13
A similar shift (<2 cm-1
) is
observed for the E2low
phonon with changing Zn mass, which is dominated by the Zn
eigenvector. Corresponding data for O-enriched ZnO nanorods are presented in
chapter 5.
It is also interesting to note the changes in the FWHM of the E2high
phonon line
with isotopic enrichment. There is significant broadening of the FWHM for this
mode, however it is not isotopic in nature. The changing FWHM for different Zn
masses results from a change in the overlap of the two-phonon density of states due
to changes in phonon frequency. Ab initio calculations performed by Serrano et al.13
show that the E2high
phonon mode is near a sharp ‘ridge’ in the two-phonon density
of states, the interaction with which results in the variation in FWHM. The overlap is
less clearly observed for the E2low
mode with no change observed in the FWHM with
changing Zn mass. Our data are consistent with the results reported by Serrano et al.
The observed shifts in the phonon vibration energies in our samples are
consistent with previous findings in the literature and, therefore, demonstrate clearly,
and independently of the SIMS and PL experiments, that the desired isotopic
enrichment of the samples has been successful.
100
Figure 4.12: Raman spectra of isotopically enriched 64
ZnO, 66
ZnO and 68
ZnO
nanorods. Inset shows the signal from the Si substrate.
Sample Mode Wavenumber
(cm-1
)
ΔWavenumber
(cm-1
)
FWHM
(cm-1
)
ΔFWHM
(cm-1
)
64ZnO E2
high 439.03 - 8.11 -
66ZnO E2
high 437.73 -1.30 6.94 -1.17
68ZnO E2
high 436.86 -2.17 5.59 -2.52
64ZnO E2
low 100.60 - 1.70 -
66ZnO E2
low 99.57 -1.03 1.70 -0.00
68ZnO E2
low 98.79 -1.81 1.73 0.03
Table 4.4: Frequencies and FWHM of the E2low
and E2high
phonons for 64
ZnO, 66
ZnO
and 68
ZnO samples.
101
4.6 Optical Quality and Enrichment - Low temperature PL
Figure 4.13: PL spectra of the 66
ZnO sample showing (a) a broad range spectrum
displaying the typical PL emission from ZnO nanorods, (b) the intense UV band
edge emission in detail, and (c) the Cu-related 2.86 eV ZPL and associated SGB
(from annealed portion). (Note that (a) is a composite of (b) and (c)).
102
Low temperature PL was carried out on all samples. In this section the
general shape of the typical PL spectrum, which was observed in all samples is
introduced using spectra recorded with the SPEX spectrometer. The shifts in energy
of the BX recombinations in the band edge region are also presented using the
natZnO,
64ZnO,
66ZnO and
68ZnO samples. Further detail and discussion on these,
and the other samples in this set, is presented in chapter 6 as part of the discussion on
the Cu-related defect at 2.86 eV.
Figure 4.13(a) shows a broad spectrum including the band edge region, as
well as the Cu-related ZPL at 2.86 eV and its associated SGB in 66
ZnO (note the
linear y-scale here).14,15
66
ZnO was chosen to illustrate the general features of these
spectra as its spectrum was representative of those obtained from the other samples
and typical of the PL spectrum obtained from ZnO nanorods reported in the
literature.16
Note that figure 4.13(a) is a composite of the spectra in figures 4.13(b)
and (c) in order to show a broad spectrum.
Figure 4.13(b) shows the band edge region of the 66
ZnO sample in detail
(note the logarithmic y-scale). The intense band edge emission due to shallow donor
bound excitons is dominant at ~3.36 eV. The FX position can also be seen, although
this was not clearly observed in all samples. The LO phonon replicas of the BX
lines are also present (BX-1LO, BX-2LO and BX-3LO), spaced at intervals of ~72
meV, the characteristic LO phonon energy of the ZnO crystal structure.17
The TES
feature is also present, as are its LO replicas (TES-1LO and TES-2LO). The intense
peaks and narrow line widths (<1 meV) of the BX lines indicate the excellent optical
quality of the nanorods as grown, as discussed further below. This is consistent with
the excellent optical characteristics of ZnO nanorods grown by similar methods
reported elsewhere and demonstrates the ability to grow isotopically enriched ZnO
materials with excellent optical quality using the straightforward carbothermal
reduction VPT method, with mg quantities of source material.18
Figure 4.13(c) shows the green band emission region for 66
ZnO. This was
recorded using a portion of the 66
ZnO sample which was annealed for ten minutes at
900 °C to increase the ZPL and green band intensity.18,19
The Cu-related ZPL at
2.86 eV is present and its associated SGB is the dominant feature. Indeed, the SGB
is the primary feature of the PL spectrum in addition to the UV band edge emission.
103
The LO phonon replicas of the ZPL are clearly seen in the SGB at intervals of ~72
meV below the ZPL energy. Some weak lines from the Hg lamp used to calibrate
the spectra are also observed. Note that the PL emission is associated with the VPT-
grown nanorods and there is not a significant emission from the underlying buffer
layer. This is illustrated in figure 4.14 which shows that the band edge PL intensity
of CBD nanorods is negligible compared to that of the VPT nanorods as recorded
under the same conditions (optimised for VPT samples, i.e. very narrow slit widths).
In fact, no band edge emission was observed in the CBD nanorods grown by the
NaOH method used in these samples. However a slight increase in intensity was
observed after the CBD sample underwent a thermal cycle similar to the VPT growth
cycle, although this was still barely detectable under these conditions. A more
detailed study of the PL emission of CBD-grown nanorods has been reported and
shows a greater band edge emission from CBD nanorods grown using the other baths
described in section 2.2.3 (HMT-based and acetate-based), although still 2-4 orders
of magnitude less intense than the emission from VPT-grown nanorods.20,21
Figure 4.14: Comparison of PL intensities of VPT and CBD nanorods.
Figure 4.15 shows representative band edge spectra from selected samples.
The dominant feature in the band edge spectra is the I9 line attributed to In22
donor
bound impurities. The I8 and I6 lines attributed to Ga and Al impurities are also
104
clearly observed.23–25
The I2 line attributed to ionised In impurities26
and the surface
exciton27
(SX) emission are also visible and labelled in the figure.
Figure 4.15: Typical PL spectra of selected enriched ZnO nanorod samples showing
the band edge region including the I9 line (spectra shifted vertically for clarity).
Table 4.5: Energies and FWHM of the I9 exciton recombination in samples with
different Zn isotopic enrichments.
Sample
Average Zn
isotopic content
(amu)
I9 energy
(eV)
I9 FWHM
(meV)
64ZnO 64 3.35629 0.44
66ZnO 66 3.35662 0.36
68ZnO 68 3.35689 0.41
64/66ZnO 65 3.35641 0.33
66/68ZnO 67 3.35682 0.33
64/68ZnO 66 3.35643 0.35
64/66/68ZnO
66 3.35603 0.35
natZnO
65.4 3.35650 0.31
105
The position of the I9 In bound exciton recombination in unenriched material
at ~3.356 eV was used to measure changes in the band edge positions with changing
Zn isotope enrichment, and the other lines followed its trend. The peak positions of
the I9 line were extracted by fitting Gaussian curves to the data. The position of the
I9 line in each sample is given in table 4.5, along with its FWHM. The I9 BX energy
increases with increasing Zn isotopic mass from 64
ZnO to 68
ZnO by 0.6 meV. The
blue shift in energy recorded in the band edge region when the Zn isotope mass is
changed (0.6 meV) is comparable to that previously reported in the literature. Tsoi
et al. report an increase of ~1.0 meV in the A-exciton band gap over this range.28
This is further verification of the successful isotopic enrichment of our samples. We
note that Manjón et al. report blue shifts in the band edge region of ~1.7 meV from
64ZnO to
68ZnO for single crystals.
29 However since Manjón et al.’s results come
from measurements of the I4 bound exciton recombination and since this emission
has been attributed to H at an O site30
(i.e. a H atom surrounded by Zn atoms), this
recombination may be more strongly affected by Zn isotopic changes (due to both
local and extended vibronic modes) and undergo a different shift than other shallow
donor bound excitons on Zn sites surrounded by O atoms (such as the In-related I9).
The very narrow FWHM of the exciton recombinations in these nanorod
samples demonstrates their excellent optical quality and therefore their suitability for
use in defect and impurity studies using PL. The I9 FWHMs here are in the range of
0.31-0.44 meV as shown in table 4.5 (comparable to very high optical quality
commercial single crystal ZnO26
). This is much narrower than the line widths
observed by others in single crystal isotopically enriched ZnO. Manjón et al. have
reported line widths of < 5 meV29
and Tsoi et al. observed BX PL features of widths
of ~2-8 meV28
in single crystal isotopically enriched ZnO samples. Given this
excellent optical quality, these samples were used to carry out a number of detailed
optical studies of defects in ZnO, including the Cu-related emission ZPL at 2.86 eV,
as well as the manifold of closely spaced near band edge I-lines due to donor bound
exciton emission.
106
4.7 FX Energies - Reflectance Spectroscopy
Reflectance spectroscopy was carried out as outlined in chapter 2. The
purpose of this was to obtain the FX energies (specifically the A-exciton) of the
isotopically enriched samples. This would complement the PL results outlined in
section 4.6. It was hoped that the same trend as seen with the BX lines of increasing
exciton recombination energy with the Zn isotopes moving from 64
Zn to 68
Zn would
be observed. The valence band of ZnO is split into three band called the A, B and C
bands, each of which has a specific FX energy which can be observed using
reflectance spectroscopy, depending on the electric field polarisation.28,31
Further
detail on the valence band splitting and the structure of the ZnO reflectance spectrum
can be found in reference 16.
Figure 4.16: Reflectance spectrum of a (a) single crystal ZnO sample showing the
A-exciton and B-exciton clearly, and (b) reflectance spectrum of the 64
ZnO, 66
ZnO
and 68
ZnO samples. A and B label the A- and B-excitons.
Figure 4.16 shows the reflectance of a sample of (a) a ZnO single crystal and
(b) representative spectra of the Zn-enriched nanorods recorded using the FT
spectrometer. The A- and B-excitons are clearly shown in (a) and are observed in
(b) also. The spectra displayed strong band edge PL emission arising from the above
band gap light emanating from the deuterium lamp which dominated the reflected
light. However it proved difficult to observe the FX energies clearly in the nanorod
107
samples and it was not possible to detect any trend is the change in energy with
different isotopic content. The spectra are noisy because the nanorod sample
morphology leads to strong scattering and absorption (with subsequent PL emission)
and little specular reflection due to the dull and diffusely reflecting surfaces in these
samples. Extensive efforts were made to limit noise by using lenses to focus the
light to a bright spot on the sample, using filtering to ensure only light from the
region of interest reached the FT spectrometer, running and averaging many scans,
careful optical alignment and other measures. These measures were ultimately
unsuccessful, however this does not undermine the evidence of successful isotopic
enrichment obtained using other methods as presented in this chapter.
4.8 Conclusions
The three-step growth process, previously developed in our group, has been
successfully adapted in order to grow nanorods of ZnO isotopically enriched with
different Zn isotopes. Samples of nat
ZnO, 64
ZnO, 66
ZnO, 68
ZnO, 64/66
ZnO, 66/68
ZnO,
64/68ZnO and
64/66/68ZnO were grown using this VS CTR-VPT techniques on CBD
buffer layers using this novel method. SEM revealed a dense coverage of vertical, c-
axis aligned nanorods over a large sample area for nearly all samples, with slight
variations seen in one sample. XRD confirmed the presence of wurtzite ZnO and
excellent nanorod alignment and crystal quality. SIMS data confirm the successful
isotopic enrichment consistently using two independent SIMS systems. Raman data
show a shift of >1 cm-1
in the peak position of the Raman scattered peaks due to the
E2low
and E2high
phonon modes when the Zn isotope is changed from 64
Zn to 68
Zn,
consistent with previous work on samples with different isotopic enrichments, again
confirming successful isotopic substitution. Low temperature PL measurements
revealed the excellent optical quality of the samples with band edge emission
emissions displaying strong and spectrally narrow BX lines in all samples. An
increase in the I9 exciton recombination energy of ~0.6 meV when the Zn isotopic
content was changed from 64
ZnO to 68
ZnO, and narrow line widths of <1 meV, were
observed. This is consistent with previous results reported for single crystals, further
confirming successful isotopic substitution of these high structural and optical
108
quality nanorods using this simple and reliable growth method. The excellent optical
quality also confirms the potential for the use of these samples in defect and impurity
studies using PL.
4.9 References
1 D. Byrne, E. McGlynn, K. Kumar, M. Biswas, M.O. Henry, and G. Hughes, Cryst.
Growth Des. 10, 2400 (2010).
2 R.B. Peterson, C.L. Fields, and B.A. Gregg, Langmuir 20, 5114 (2004).
3 L.E. Greene, M. Law, D.H. Tan, M. Montano, J. Goldberger, G. Somorjai, and P.
Yang, Nano Lett. 5, 1231 (2005).
4 R.B. Saunders, S. Garry, D. Byrne, M.O. Henry, and E. McGlynn, Cryst. Growth
Des. 12, 5972 (2012).
5 E. McCarthy, R.T. Rajendra Kumar, B. Doggett, S. Chakrabarti, R.J. O’Haire, S.B.
Newcomb, J.-P. Mosnier, M.O. Henry, and E. McGlynn, J. Phys. D. Appl. Phys. 44,
375401 (2011).
6 R.T.R. Kumar, E. McGlynn, M. Biswas, R. Saunders, G. Trolliard, B. Soulestin, J.-
R. Duclere, J.P. Mosnier, and M.O. Henry, J. Appl. Phys. 104, 084309 (2008).
7 B.-H. Hwang, J. Phys. D. Appl. Phys. 34, 2469 (2001).
8 S. Desgreniers, Phys. Rev. B 58, 14102 (1998).
9 W.C. O’Mara, R.B. Herring, and L.P. Hunt, editors , Handbook of Semiconductor
Silicon Technology (Noyes Publications, Park Ridge, New Jersey, USA, 1990).
10 V. Mote, Y. Purushotham, and B. Dole, J. Theor. Appl. Phys. 6, 6 (2012).
11 K. Ghule, A.V. Ghule, B.-J. Chen, and Y.-C. Ling, Green Chem. 8, 1034 (2006).
12 J. Serrano, F. Widulle, A.H. Romero, A. Rubio, R. Lauck, and M. Cardona, Phys.
Status Solidi 235, 260 (2003).
13 J. Serrano, F. Manjón, A. Romero, F. Widulle, R. Lauck, and M. Cardona, Phys.
Rev. Lett. 90, 055510 (2003).
14 R. Dingle, Phys. Rev. Lett. 23, 579 (1969).
109
15 D. Byrne, F. Herklotz, M.O. Henry, and E. McGlynn, J. Physics. Condens. Matter
24, 215802 (2012).
16 C.F. Klingshirn, B.K. Meyer, A. Waag, A. Hoffmann, and J. Geurts, ZnO: From
Fundamental Properties Towards Novel Applications (Springer-Verlag, Berlin,
Heidelberg, 2010).
17 L. Wang and N.C. Giles, J. Appl. Phys. 94, 973 (2003).
18 G. Xing, G. Xing, M. Li, E.J. Sie, D. Wang, A. Sulistio, Q. Ye, C.H.A. Huan, T.
Wu, and T.C. Sum, Appl. Phys. Lett. 98, 102105 (2011).
19 N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, and G. Cantwell,
Appl. Phys. Lett. 81, 622 (2002).
20 D. Byrne, The Growth and Characterisation of Ordered Arrays of Zinc Oxide
Nanostructures and Optical Studies of Defects in Zinc Oxide, Dublin City
University, 2012.
21 D. Byrne, E. McGlynn, J. Cullen, and M.O. Henry, Nanoscale 3, 1675 (2011).
22 S. Muller, D. Stichtenoth, M. Uhrmacher, H. Hofsass, C. Ronning, and J. Roder,
Appl. Phys. Lett. 90, 012107 (2007).
23 K. Johnston, M.O. Henry, D. McCabe, E. McGlynn, M. Dietrich, E. Alves, and M.
Xia, Phys. Rev. B 73, 165212 (2006).
24 M. Schilling, R. Helbig, and G. Pensl, J. Lumin. 33, 201 (1985).
25 B.K. Meyer, H. Alves, D.M. Hofmann, W. Kriegseis, D. Forster, F. Bertram, J.
Christen, A. Hoffmann, M. Straßburg, M. Dworzak, U. Haboeck, and A. V. Rodina,
Phys. Status Solidi 241, 231 (2004).
26 J. Cullen, D. Byrne, K. Johnston, E. McGlynn, and M.O. Henry, Appl. Phys. Lett.
102, 192110 (2013).
27 M. Biswas, Y.S. Jung, H.K. Kim, K. Kumar, G.J. Hughes, S. Newcomb, M.O.
Henry, and E. McGlynn, Phys. Rev. B 83, 235320 (2011).
28 S. Tsoi, X. Lu, A.K. Ramdas, H. Alawadhi, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 74, 165203 (2006).
29 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
Cardona, Solid State Commun. 128, 35 (2003).
110
30 E. Lavrov, F. Herklotz, and J. Weber, Phys. Rev. B 79, 165210 (2009).
31 H. Alawadhi, S. Tsoi, X. Lu, A.K. Ramdas, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 75, 205207 (2007).
111
Chapter 5: O Isotope-enriched ZnO
Nanorods
5.1 Introduction
This chapter presents the results of the growth of ZnO nanorods enriched with
the 18
O isotope. These samples were produced using two separate methods as
described in section 2.3. The first was a VS method carried out by modifying the
CTR-VPT setup used in the growth of the Zn-enriched samples. The second was a
VLS technique carried out in collaboration with colleagues in the Institute for Solid
State Physics in the University of Jena, Germany. Specifically, this second set of
growths was carried out by Mr. Lukas Trefflich and assisted by the author in the
group of Prof. Carsten Ronning during a research visit by the author in September
2015, using isotopically enriched source powders produced beforehand in DCU by
the author. The SEM images of these samples presented in section 5.2 were obtained
during this visit by Mr. Trefflich and assisted by the author. Both sets of samples
were characterised using SEM, XRD, SIMS, Raman spectroscopy and low
temperature PL. In this chapter, some initial, unsuccessful, attempts to produce O-
enriched nanostructures are described, followed by the results of the two successful
methods. Table 5.1 presents a summary of the O-enriched ZnO nanorods
successfully produced by the two methods, including those with natural Zn16
O, and
112
nominal Zn18
O and Zn16/18
O, compositions. They are distinguished by VS or VLS
labels to indicate the growth method.
Label Sample composition Description
Zn16
O-VS Zn16
O Method 1: Modified VS CTR-VPT
(in DCU) Zn
16/18O-VS Zn
16/18O
Zn18
O-VS Zn18
O
Zn16
O-VLS Zn16
O
Method 2: VLS VPT (in Jena) Zn16/18
O-VLS Zn16/18
O
Zn18
O-VLS Zn18
O
Table 5.1: Sets of O-enriched ZnO nanorods successfully produced using both
methods.
5.1.1 Initial Experiments regarding O-enriched Material
The process of devising a method to grow O-enriched ZnO nanorods proved
significantly more difficult than for the Zn-enriched material. A number of attempts
were made using the VPT furnace, however ultimately these attempts did not
produce deposits of the nanorod morphology desired. More importantly, these
deposits did not display the very high optical quality required for detailed PL studies
of defects. These unsuccessful efforts are described briefly in this section in order to
give some sense of the challenges involved. However, the rest of this chapter is
devoted to the characterisation of samples produced using the two ultimately
successful methods.
The central challenge confronting all these experiments is that we wished to
be able to control the amount of 18
O2 gas used (since this is an expensive material)
and also to be able to remove as much as possible the ubiquitous 16
O when we
wished to grow fully 18
O enriched material.
For this reason we looked firstly at a few methods centred on direct oxidation
of Zn material, for which approach some recipes are available in the literature. The
113
tests of these growth recipes were all actually conducted with natural O in the first
instance, to determine if nanostructured material of suitable morphology and optical
quality would be obtained. We would naturally only have proceeded to the use of
18O2 in the event of successful growths with natural O. The first tests involved
attempting to deposit some ZnO by heating Zn powder at temperatures around ~500-
600°C, using CBD buffers or Au-coated Si, both in Ar flows and in oxygen
atmospheres at low pressures. In some cases a slow temperature ramp was also used.
Some of these experiments resulted in ZnO deposits on the inside of the quartz tube,
but none resulted in nanostructure deposits on the substrates. This was followed by
an attempt at growing nanowires from the Zn powder simply deposited on bare Si
substrates by direct oxidation similar to reports in the literature.1,2
However, the
nanorods obtained by this method were of very poor optical quality. Another series
of experiments using CBD buffer layers or Au-coated Si, a temperature of ~600-
900°C and an oxygen atmosphere of ~10 mbar was carried out. These resulted in
ZnO deposits on the tube and placing the samples at different locations in the tube
resulted in ZnO deposits, but again these were not nanostructured and were of poor
optical quality.
We then decided that direct oxidation methods seemed rather unpromising,
so experiments based on carrying out the normal VPT procedure, but with extra steps
involving evacuation and backfilling the tube with an 21:79 O2:N2 mix, to simulate
an air atmosphere, but using a smaller diameter tube were also unsuccessful. This
approach was intended to allow us to control the O isotope in the artificial air
atmosphere, but also to use a smaller amount of 18
O2 in the smaller tube. In this
case, although the Ar flow rate was adjusted proportionately to the tube cross
sectional area, no deposits were obtained. This demonstrates just how sensitive the
VPT process is to any change in parameters. A series of experiments were then
carried out based on previous work in our group involving controlling the amount of
oxygen in the tube with another MFC along with the Ar flow.3 However these also
did not lead to the desired nanorod growth, perhaps again due to changing
parameters such as the tube size and oxygen flow, in order to minimise future
enriched O2 gas usage. Finally, direct sublimation based on the work of Prof.
Carsten Ronning’s group4 was attempted but was not successful due to our VPT
furnace being limited in temperature to 1100°C.
114
While these experiments were not successful in achieving the desired ZnO
nanorods of high structural and optical quality, they were important as they formed
part of the process of guiding our thinking in terms of developing the growth
methods which ultimately were successful. The rest of this chapter is devoted to
those methods.
5.1.2 Method 1: Modified VS CTR-VPT method
As described in sections 2.3.1 and 4.1, the CTR-VPT growth used for the Zn-
enriched growths is based on the reduction of ZnO powder by the graphite to
produce Zn vapour and carbon monoxide (CO) by equation 4.1. The Zn vapour is
then re-oxidised using the residual oxygen in the tube (rather than the O initially in
the ZnO powder, which is captured by C to form CO) in a VS process at the
energetically favourable sites provided by the CBD layer produced using the NaOH-
based reaction. The work described in section 5.1.1 above in terms of evacuating
and backfilling a smaller tube with a 21:79 O2:N2 mix, to simulate an air atmosphere,
but now using the normal CTR-VPT tube size was the natural next step. When this
method was tried O-enriched ZnO nanorods were successfully produced by
removing all the residual atmospheric O2 from the tube by evacuation to 1 mbar, and
then reintroducing O2 gas with the desired isotopic content, and in the same amounts
as typical atmosphere (~21%), along with N2 gas (79%), to a total starting pressure
of 1 atmosphere and carrying out the VPT growth as before. This method had the
disadvantage of requiring significant quantities of isotopically enriched 18
O2 gas, but
was otherwise very successful. The samples produced with this procedure are
labelled ‘VS’ as in table 5.1. In order to try and confirm that only the residual O2,
which can be evacuated and replaced with 18
O2, contributes to the O in the ZnO
nanorods, growth of a sample was attempted by this method in a nitrogen
atmosphere only. This would confirm that there is no contribution of O from the
ZnO source material, or from older deposited ZnO material on the tube from
previous growths. Figure 5.1 shows an SEM image of this sample. The CBD buffer
layer appears slightly cracked, perhaps due to thermal expansion at the VPT
temperatures, and shows some grain alteration, but there is no evidence of VPT
115
nanowire growth and thus it was confirmed that no VPT nanorod growth occurred
under these conditions. This is consistent with previous work in our group that
found that the growth process is quenched when the residual O2 in the tube is
sufficiently depleted by a longer Ar flow during the initial steps of the CTR-VPT
process, e.g. after about 15-30 minutes Ar flow.3,5
Figure 5.1: Image of CBD buffer following VPT step in oxygen-depleted atmosphere.
No nanorod growth is observed.
5.1.3 Method 2: VLS VPT method
We were aware that certain groups (including our collaborators in Jena) use a
direct sublimation of ZnO powders to grow ZnO nanostructures (and that the O in
the nanostructures comes from the initial powder). Given that we had developed
some knowledge in how to directly oxidise Zn to create ZnO material (as described
in section 5.1.1 above for the unsuccessful attempts) we decided that we could use a
direct oxidation of Zn metal to “trap” the 18
O to form a Zn18
O powder, which could
then be used in the VLS VPT method. Having produced such powders we decided to
go to our Jena collaborators to actually perform the VLS VPT growths, since our
own furnace does not reach the required temperatures, but more importantly since
they have the necessary experience in producing samples with such a method, and
116
establishment of a new growth method in a new location is quite a time-consuming
task, even if one has the recipe, because of the sensitivity of such high temperature
methods to a range of parameters, as commented on previously. Because we had
effectively “trapped” the 18
O to form a Zn18
O powder, we could transport the source
material quite easily.
For this VLS method, as described in section 2.3.2, source powders of ZnO
and isotopically enriched Zn18
O were produced in DCU by oxidising Zn metal
powder. The samples were grown on substrates coated with Au. The Au coating
melts during growth and forms droplets on the substrate which act as the
energetically favourable nucleation sites. The nanorods in this case are not
preferentially aligned as the ZnO is not epitaxially matched to Si/SiO2 and there is no
textured CBD buffer layer in this case. The samples produced with this procedure
are labelled ‘VLS’ as in table 5.1.
5.2 Morphology – SEM
Figure 5.2: SEM images of sample Zn16
O-VS at (a) ~40° and (b) ~90°, and the
middle of sample Zn16/18
O-VS at (c) ~40° and (d) ~90°.
117
Typical images of the morphology of the VS samples grown by method 1 are
shown in figure 5.2. Figures 5.2(a) and (b) shows the morphology observed in the
Zn16
O-VS sample. This is also typical of the morphology observed in the Zn18
O-VS
sample, as well as near the edges of the Zn16/18
O-VS sample. The morphology is
similar to that shown in previously in figure 4.5(d), where the nanorods are on the
order of 10 μm in length. They have become entangled as described previously in
chapter 4. Figures 5.2(c) and (d) are from the middle part of the Zn16/18
O-VS
sample. In this case, the nanorods are the more typical 2-3 μm in length and are
more clearly vertically aligned. In this case, the length of the nanorods increased,
moving from the middle of the sample to the edge. It is interesting to note that for
this sample a slightly longer Ar flush time was used than the other VS samples of
about 6-7 minutes instead of 5 minutes. This is closer to the usual flush time as used
in the Zn-enriched set. This would result in a smaller amount of residual O2 in the
tube as the temperature was increased and therefore shorter nanorods due to the
growth ceasing when the residual oxygen levels are depleted. This provides a
demonstration of just how extremely sensitive the VPT process is to the various
parameters, in this case the residual oxygen content during the growth phase.
Overall, the dimensions of these nanorods are in line with those previously described
for the Zn-enriched samples in chapter 4. Coverage on these samples was excellent,
with the entire sample coated with well aligned ZnO nanorods with high density and
structural quality.
Figure 5.3 displays examples of the typical morphology observed using the
SEM system in Jena of the VLS samples grown there using method 2. Of the 7-8
pieces of Au-coated Si placed in the furnace for these growths, ZnO nanorods were
deposited typically on the middle three or four pieces, which were ~15 cm from the
middle of the ZnO source powder. The temperature of the tube here was lower than
at the centre of the furnace where the source powder was located, around ~950-
1150°C4, and this appears to be a key parameter in whether deposits were observed
on a particular piece, as the pieces were spread over a few cm.
118
Figure 5.3: SEM images of the morphology of the VLS samples with 30° tilt; (a)
Zn16
O-VLS, (b) Zn16
O-VLS showing hexagonal structure, (c) Zn18
O-VLS and (d)
Zn16/18
O-VLS.
Figure 5.3(a) shows an example from the Zn16
O-VLS sample of the typical
morphology observed on all the substrates with deposits across all three isotopic
compositions as listed in table 5.1. The nanorods are generally longer than seen in
the VS O- or Zn-enriched samples, with lengths of several tens of μm observed in all
samples. The diameters of several individual nanorods were measured and were in
the range of ~100-220 nm, with most being between 120-150 nm. These dimensions
agree with the ranges previously reported using this growth technique.4,6
Figure
5.3(b) shows a close up of the end of a nanorod of Zn16
O-VLS clearly showing the
hexagonal crystal structure associate with wurtzite material. Figures 5.3(c) and (d)
are from the Zn18
O-VLS and Zn16/18
O-VLS samples respectively, and display similar
morphology. The greatest difference in morphology between these samples and
those grown with method 1 is that there is no preferential alignment of the nanorods
along the c-axis. This is because of the lack of an epitaxially matched c-axis
matched buffer layer in this case. Coverage on the substrates with deposits was
excellent, with the entire surface containing these high density nanorods of high
structural quality.
119
5.3 Alignment and Crystal Quality – XRD
Figure 5.4: 2θ-ω diffractograms of the VS O-enriched samples dominated by the Si
substrate peak at 69.1˚ and the ZnO peak at 34.4˚. Inset shows rocking curves of the
ZnO (0002) peak.
Figure 5.4 shows the 2θ-ω XRD scans of the VS samples grown with method
1. In this case, the diffractograms are similar to those of the Zn-enriched samples
presented in section 4.3. The diffractograms are again dominated by the ZnO (0002)
reflection at ~34.4° and the Si (004) substrate reflection at ~69.1°. The second order
ZnO (0004) reflection and the forbidden Si (002) reflection7 are present similar to
the Zn-enriched samples, as also are some features assigned to the backing tape used
to mount the samples and contamination from the x-ray tube8, marked T and X
respectively. Table 5.2 lists the peaks position of the ZnO (0002) reflection and their
FWHM. This strong reflection indicates successful deposition of wurtzite ZnO on
the substrate. The narrow values of FWHM (<0.2°) indicate high crystal quality in
terms of out-of-plane coherence length. The inset shows the rocking curves of the
120
ZnO (0002) reflection. Their peak positions and FWHM are listed in table 5.3. The
high quality of the ZnO crystal structure, and the high degree of preferential
alignment along the c-axis, is demonstrated by the narrow line widths of the rocking
curves (1.96-3.08°). These data are again consistent with the SEM images of these
samples presented in section 5.2. The peak position of the Zn16
O-VS rocking curve
is slightly shifted due to a variation in the tilt of the samples as mounted on the
instrument stage.9
Table 5.2: ZnO (0002) 2θ peak and FWHM, and calculated crystallite size and c
lattice constant for the VS and VLS O-enriched samples.
Table 5.3: Peaks and FWHM of the rocking curves of the 34.4˚ ZnO (0002)
reflection the VS O-enriched samples.
Figure 5.5 shows the corresponding 2θ-ω diffractograms for the VLS
samples grown using method 2. The diffractograms have a distinctly different
appearance to those previously observed in the VS samples and the Zn-enriched
ones. The diffractograms are dominated by the Si (004) reflection of the substrate,
Sample ZnO (0002)
peak 2θ (°)
ZnO (0002)
FWHM 2θ
(°)
Crystalline
size (nm)
ZnO c lattice
constant
(nm)
Zn16
O-VS 34.41 0.199 48 0.521
Zn16/18
O-VS 34.45 0.189 52 0.521
Zn18
O-VS 34.45 0.197 49 0.521
Zn16
O-VLS 34.44 0.203 47 0.521
Zn16/18
O-VLS 34.41 0.212 44 0.521
Zn18
O-VLS 34.42 0.192 51 0.521
Sample Rocking curve peak (°) FWHM (°)
Zn16
O-VS 16.00 1.96
Zn16/18
O-VS 17.13 3.08
Zn18
O-VS 17.13 2.86
121
and the Si (002) reflection is again present. However, instead of being dominated by
the ZnO (0002) reflection at 34.4° indicating c-axis aligned wurtzite ZnO, there are a
number of peaks associated with different planes in wurtzite ZnO present. These
include the c-plane (0002), the m-plane (10-10), the s-plane (10-11), the r-plane (10-
12), the a-plane (11-20) and the (10-13) reflection. (These assignments are made
using JCPDS card number 65-3411.10
) The presence of such a number of different
ZnO planes in the 2θ-ω spectrum indicates that the material is not preferentially
aligned. Each nanorod is in a different orientation to their neighbours, and so there
are some nanorods in the appropriate orientation for each of these planes to appear as
their reflections satisfy the Bragg equation, i.e. an approximation to the powder
pattern is seen. These measurements are consistent with the SEM images in section
5.2 for these samples which show nanorods have grown all directions without any
preferential alignment. They are also consistent with the VLS growth method used
which made use of Au catalyst and not a buffer layer of c-axis aligned nanorods.
The peak positions of the ZnO (0002) reflection, and their FWHM, are shown in
table 5.2. Again the FWHM (~0.2) indicated a high quality crystal structure has
been deposited using this growth method. The inset of figure 5.5 shows the rocking
curves of these reflections. However, in this case, the rocking curves have very large
FWHM and are weaker due to the unaligned nature of the growth, as expected and in
agreement with the 2θ-ω diffractograms generally and the SEM images of these
samples.
The out-of-plane nanocrystallite sizes and the c lattice constants were
calculated using the Scherrer relation (eqn. 4.2)11
and the Bragg formula (eqn. 2.1)
respectively for both the VS and VLS samples and are also shown in table 5.2. The
crystalline sizes (44-52 nm) are similar to those calculated for the Zn-enriched
samples, and the c lattice constant is at the expected value of 0.521 nm in all
samples.12
The error is ±7 nm in the nanocrystallite size and ±0.004 nm in the lattice
constant.
Overall, these data indicate that ZnO nanorods of high quality crystal
structure have been successfully produced using both the VS and VLS methods.
Preferential alignment along the c-axis is displayed in the VS samples, but not in the
VLS ones, in agreement with the SEM data and expectations arising from the growth
methods used.
122
Figure 5.5: 2θ-ω diffractograms of the VLS O-enriched samples dominated by the Si
substrate peak at 69.1˚ and containing multiple ZnO peaks. Inset shows rocking
curves of the ZnO (0002) peak.
5.3.1 Note on Strain
The state of strain in a crystal can affect both the PL emission line energies as
well as (though to a lesser extent) Raman line positions and this has relevance to
some of our further characterisation measurements. The 2θ-ω diffractograms above
provide information concerning the crystal state and strain in a direction
perpendicular to the substrate, i.e. out-of-plane (due to the symmetric nature of the
reflections), because this method essentially maps along a line in reciprocal space
normal to the substrate. XRD also allows us to examine the sample crystallinity and
strain along directions which are not perpendicular to the substrate (i.e. with an in-
plane component) by mapping along a line in reciprocal space which is not normal to
the substrate. Since the nanorods produced here are essentially ‘free standing’
structures (i.e. with nanorod sidewalls unconstrained by neighbouring material) on
ZnO buffer layers, we expect that strain in these samples would be minimal, as each
individual nanorod is essentially a small isolated single crystal in terms of structure.
123
The 2θ-ω diffractograms above confirm that we see no evidence of out-of-plane
strain at the resolution of our XRD instrument. Using our scan step size of for these
out-of-plane scans of 0.05°, this leads to a minimum detectable strain of 0.00145,
and stress of 0.3045 GPa. In order to check for in-plane strain, the 2θ positions of a
number of other planes in the c-axis aligned VS samples were measured by tilting
the sample to suitable angles and then measuring the 2θ-ω diffractograms but with
asymmetric reflections, which enables the mapping along a line in reciprocal space
which is not normal to the substrate. Specifically, the s-plane and r-plane were
selected for this measurement as they are among the stronger reflections. To access
these reflections in the c-axis aligned nanorods, the samples were tilted in the XRD
apparatus by the angle between these planes and the c-axis. This essentially aligned
these planes with the x-ray source and detector as if they were vertical like the c-
plane was previously. Figure 5.6 illustrates the general physics of this in a diagram,
although the actual experimental setup was slightly different. Case (a) is the normal
XRD set up with the c-planes parallel to the substrate satisfying Bragg’s Law with
symmetric reflections around the c-axis normal to the substrate. In case (b) the
sample it tilted so that a different plane experiences the symmetric reflections to
satisfy Bragg’s Law, however now the reflections around the normal to the substrate
are asymmetric, illustrated by angles α and β. Finally, (c) shows the effect of this in
reciprocal space, showing the mapping along a line which is not normal to the
substrate.
Figure 5.6: (a) Symmetric and (b) asymmetric XRD reflections, which enables (c) the
mapping along a line in reciprocal space which is not normal to the substrate.
124
The angles of tilt needed were calculated using the formula for the angle
between planes (hkl)1 and (hkl)2 in a hexagonal structure:
cos 𝜌 =ℎ1ℎ2 + 𝑘1𝑘2 +
12
(ℎ1𝑘2 + ℎ2𝑘1) + (3𝑎2
4𝑐2) 𝑙1𝑙2
√{[ℎ12 + 𝑘1
2 + ℎ1𝑘1 + (3𝑎2
4𝑐2) 𝑙12] [ℎ2
2 + 𝑘22 + ℎ2𝑘2 + (
3𝑎2
4𝑐2) 𝑙22]}
Eqn. 5.1
where h, k and l correspond to each plane using Miller notation, a and c are the ZnO
lattice constants and ρ is the angle between planes. The angles of tilt calculated were
61.6103° for the s-plane and 42.7728° for the r-plane. A 2θ-ω scan was then
performed in each case. Strain in the nanorods can be expected to appear in the
spectra as a shift in the 2θ peak position as the peak angle and inter-planar distance
are linked through the Bragg equation.13
Figure 5.7: s-plane and r-plane 2θ peaks (with in-plane component) in the tilted VS
O-enriched samples. The green vertical lines represent the calculated reference
values from the JCPDS card number 65-341110
.
125
Table 5.4: s-plane and r-plane 2θ peaks in the VS (with in-plane component) and
VLS O-enriched samples (from standard out-of-plane diffractograms for
comparison).
Figure 5.7 displays the s-plane and r-plane reflections with in-plane
components in the c-axis aligned VS samples. The green lines represent the
calculated reference values of 36.256° and 47.541° respectively (JCPDS card
number 65-341110
) and the values observed are shown in table 5.4. Table 5.4 also
includes the out-of-plane peak positions of these reflections in the non-aligned VLS
samples for comparison. The peaks recorded for the VS samples in these out-of-
plane scans were quite weak and noisier than the normal 2θ-ω scans due to the lack
of epitaxial in-plane alignment of the nanowires meaning that only a fraction
matched the diffraction condition for any azimuthal angular setting, and were also
broader since the coherence length is geometrically shortened compared to the out-
of-plane value. The peaks have very good agreement with the expected values, to
within 0.01° in all but one case. The Zn16/18
O-VS sample had the largest change,
although this was still limited to ~0.03° which is within the resolution limit of our
XRD system. These data show that there is no detectable strain in these samples
using our XRD system, which may cause shifts in the Raman phonon energies and
PL band edge peaks, and that supports the other measurements indicating very high
structural quality.
Sample
ZnO s-plane
(10-10) peak
2θ (°)
ZnO r-plane
(10-12) peak
2θ (°)
Zn16
O-VS 36.253 36.544
Zn16/18
O-VS 36.253 36.574
Zn18
O-VS 36.248 36.543
Zn16
O-VLS 36.258 47.543
Zn16/18
O-VLS 36.230 47.530
Zn18
O-VLS 36.244 41.549
126
Note that between the submission of this thesis for examination and the viva-
voce examination, the XRD strain measurements with an in-plane component
described here for the VS O-enriched samples were carried out on the Zn-enriched
samples and the VLS O-enriched samples, in order to complete this study on the full
set of samples presented in this work. These data are presented in appendix C and
ultimately indicate no discernible strain in agreement with this section.
5.4 Isotopic Enrichment – miniSIMS
Figure 5.8: Typical miniSIMS spectra of the Zn16
O and Zn18
O samples in the VS set
(a and b) and the VLS set (b and d).
While SIMS and miniSIMS measurements carried out on the Zn-enriched
samples (section 4.4) where successful in confirming the high purity of isotopic
enrichment in those samples, such measurements of the O-enriched samples proved
127
to be less successful. These were carried out using the miniSIMS system only as the
SIMS system did not operate with the negative mode needed to measure negative
ions such as O-. Figure 5.8 shows typical miniSIMS spectra recorded for the (a)
Zn16
O-VS, (b) Zn18
O-VS, (c) Zn16
O-VLS, and (d) Zn18
O-VLS samples. For the
Zn16
O samples, (a and c), there is a large peak at 16 amu which is attributed to O-
ions, and nothing is observed at 18 amu. This indicated the presence of 16
O as was
expected for these samples. The peak at 17 amu is attributed to OH- ions.
When testing the Zn18
O samples, it expected that these two peaks would shift
to 18 amu and 19 amu respectively, and therefore indicate the presence of 18
O- and
18OH
- ions. This would be a nice confirmation of successful oxygen enrichment,
analogous to the confirmation of Zn enrichment in those samples. However, as seen
in figure 5.8(b and d), the Zn18
O only display a very small peak at 18 amu and the
large peaks at 16 amu and 17 amu remain. While this shows the presence of 18
O as
expected, it does not give direct confirmation of enrichment because of the continued
presence of 16
O. It is strongly suspected that, since the samples are exposed to air,
that adsorbed oxygen compounds on the samples lead to this result. The unenriched
buffer layer in the VS samples, and the vacuum of only ~10-6
mbar in this system,
could lead to the continued presence of natural oxygen and its compounds thus
preventing a direct measurement of O-enrichment. It is also notable that one might
expect to see a peak at 18 amu in any case due to H2O molecules adsorbed in the
sample or in the chamber, however this was not observed in these measurements as it
is a positive ion.
A number of attempts were made to isolate the Zn18
O nanorods from such
contamination. This included (i) using the depth profile mode to ‘dig’ into the
sample and bypass surface contamination, (ii) imprinting the nanorods on a clean
piece of Au wire which had been baked in vacuum by lightly pressing the sample
face down in the Au surface, (iii) imprinting some nanorods on a fresh piece of Si
which had its native oxide layer removed by etching with hydrofluoric acid. Despite
these and other extensive attempts, we were unable to isolate the Zn18
O nanorods
from such 16
O contamination and all other spectra recorded resembled those in figure
5.8. However, even without such direct confirmation in this case, the Raman and PL
data presented later in this chapter provide compelling evidence that O-enrichment
has been successful in both the VS and VLS samples.
128
5.5 Phonon Frequencies and Enrichment – Raman
Raman spectroscopy was performed on both VS and VLS O-enriched samples,
focussing on the E2high
and E2low
phonon modes as was done with the Zn-enriched
samples in section 4.5, in DCU with the assistance of Dr. Rajani Vijayaraghavan.
Figure 5.9: Raman spectra of the VS method O-enriched ZnO nanorods. Inset shows
the signal from the Si substrate. *Orange line is the second scan of the Zn16/18
O-VS
sample in middle of sample in region with shorter nanorods.
Figure 5.10: Raman spectra of the VLS method O-enriched ZnO nanorods. Inset
shows the signal from the Si substrate.
129
Figures 5.9 and 5.10 show these spectra for the VS and VLS sets
respectively. Table 5.5 gives the line phonon frequencies, their FWHM, and the
differences in these from the Zn16
O sample in each case. Note that in the case of the
Zn16/18O
-VS sample, Raman was performed both at the edge (where the morphology)
matched the other samples in this set, and in the middle (where the nanorods were
shorter as shown figure 5.2). This extra scan in shown in orange in figure 5.9 and
marked with * in table 5.5.
Changing the isotope from 16
O to 18
O, the E2high
mode induces a shift of
16.51 cm-1
in the VS set and 8.42 cm-1
in the VLS set. These figures are much larger
than the 2.17 cm-1
shift observed when changing the Zn isotope from 64
Zn to 68
Zn.
This is not surprising as this mode is dominated by the O atom vibration and
therefore we would expect a larger change when changing the O isotope than the Zn.
Shifts of ~21 cm-1
have been reported by Serrano et al. at low temperatures14
and~23
cm-1
at room temperatures.15
The widths of these lines, in the samples with 18
O
included, are much larger than those observed for Zn enrichment, in some cases
three times greater. Serrano et al. also observed the increase in line widths, and
attributes the change in line widths in the isotopically pure samples to the ‘ridge’ in
the two-phonon density of states in this region as discussed in section 4.5. The
changing O masses shift the phonon frequencies along this ridge, which itself only
depends on the Zn mass in the region of E2high
which they studied. There is an
additional broadening effect for the mixed Zn16/18
O samples, which they attribute to
additional elastic scattering of phonons due to the O isotopic variations and a
reduction in phonon lifetime as a result, and they observed this sample to have the
widest line. It is noted that although the broad trends are in agreement with these
reports, the widths recorded here are larger than those reported by Serrano et al, who
observed an increase in the widths of the these lines from 6 cm-1
to 11 cm-1
moving
from Zn16
O to Zn18
O at room temperature.15
We observe such changes from 6 cm-1
to 16 cm-1
, and 6 cm-1
to 18 cm-1
, in the VS and VLS samples respectively. In
addition, for the VLS samples, the Zn18
O-VLS sample is the widest, not the
Zn16/18
O-VLS sample. It is also of note than the peak position of the middle and
edge parts of Zn16/18
O-VS don’t match, although we do not think morphology
changes should have this effect.
130
The E2low
mode however experiences very little change as this mode is Zn-
dominated. The shifts of 0.37 cm-1
and 0.14 cm-1
are much smaller than the 1.81 cm-
1 shift observed for this mode in the Zn-enriched samples. This mode also remains
quite narrow in width and has only small changes in width, although larger than the
changes in width seen with Zn isotope changes.
Sample Mode Wavenumber
(cm-1
)
ΔWavenumber
(cm-1
)
FWHM
(cm-1
)
ΔFWHM
(cm-1
)
VS samples
Zn16
O-VS E2high
436.85 - 6.73 -
Zn16/18
O-VS E2high
426.25 -10.6 20 13.27
Zn16/18
O-VS (*) E2high
430.43 -6.42 21.9 15.17
Zn18
O-VS E2high
420.34 -16.51 16.37 9.64
Zn16
O-VS E2low
97.51 - 2.1 -
Zn16/18
O-VS E2low
97.52 0.01 1.93 -0.17
Zn16/18
O-VS (*) E2low
97.59 0.08 1.92 -0.18
Zn18
O-VS E2low
97.14 -0.37 1.96 -0.14
VLS samples
Zn16
O-VLS E2high
437.32 - 6.35 -
Zn16/18
O-VLS E2high
432.07 -5.25 16.27 9.92
Zn18
O-VLS E2high
428.9 -8.42 18.54 12.19
Zn16
O-VLS E2low
97.71 - 1.76 -
Zn16/18
O-VLS E2low
97.54 -0.17 2.32 0.56
Zn18
O-VLS E2low
97.57 -0.14 1.81 0.05
Table 5.5: Frequencies and FWHM of the E2low
and E2high
phonons for the VS and
VLS O-enriched samples. *Second scan of the Zn16/18
O-VS sample in middle of
sample in region with shorter nanorods.
While the Raman spectra indicate that O-isotopic enrichment has been
successful in both sets, it appears to indicate a lower level of enrichment than
expected, and additionally that the VLS samples have been enriched to a lesser
131
extent than the VS samples, resulting in smaller E2high
mode Raman shifts. This is
consistent with the PL data presented in the next section. The wider line widths in
our samples, and in particular our Zn18
O-VLS sample which had a wider line than
the Zn16/18
O-VLS samples, also indicate that O-enrichment may be lower than
expected. Further discussion on the Raman and PL data regarding the enrichment
levels is presented in chapter 6.
5.6 Optical Quality and Enrichment - Low temperature PL
Low temperature PL was carried out on all samples in both the VS and VLS
set. The PL data is briefly introduced here using spectra recorded on the SPEX
system and a more complete treatment, including all the PL data and the study of the
Cu-related defect at 2.86 eV in presented in chapter 6.
Figure 5.11: Effect of the 900°C 10 minute anneal on the green band region. (Note
the unstructured spectrum has been multiplied x10 for clarity. The intense, sharp
lines are from the Hg lamp used for calibration.)
The Cu-related SGB was not present in the spectra of the as prepared VS
samples. All three of these samples were annealed at 900°C for ten minutes to
132
activate the SGB.16,17
The only other difference in this case to the Zn-enriched
samples which were annealed in this way is that the O-enriched samples were
annealed in the atmosphere in which they were grown, i.e. ~79% N2, and ~21% 16
O2,
18O2 or mixed
16/18O2. The effect of this anneal on the green band is shown in figure
5.11 using Zn16
O-VS as an example. The SGB is present afterwards, displaying the
2.86 eV ZPL and LO replicas, with only a weak unstructured green band present in
the as grown samples.
The features observed in both the VS and VLS samples were similar to those
described in section 4.6 for the Zn-enriched samples. The spectra contain a strong
band edge emission and a wide SGB (after anneal in the VS samples). The LO
replicas are also present. Figure 5.12 shows the band edge spectra in detail for both
(a) the VS samples and (b) the VLS samples. The main features are once again the
In I9 line and the Al I6 line18,19
. Other prominent features which appear include the
SX feature20
which is present in some of the VS samples and the ionised In I2 line21
and ionised Al I0 line22
which are present in the VLS samples. The position of the I9
In BX recombination in unenriched material at ~3.356 eV was once again used to
measure changes in the band edge positions, in this case with changing O isotope
enrichment, and the other lines followed its trend. The positions of the I9 line in each
sample is given in table 5.6, along with their FWHM. There is an increase in energy
as the O isotopic mass changes from natural 16
O to 18
O. This blue shift is ~3.24 meV
in the VS sample set, and ~2.11 meV for the VLS samples. The changing positions
of the band edge BX lines are strong evidence of successful isotopic enrichment of
the samples. Manjón et al. have reported shifts of the BX recombinations of ~6 meV
over this range.23
In addition, Tsoi et al. have reported a similar shift of ~6 meV in
the A-exciton band gap over this range.24
Both of these measurements used single
crystal samples.
133
Figure 5.12: Band edge region of the PL spectrum in the (a) VS and (b) VLS O-
enriched samples. (Spectra offset vertically for clarity.)
Sample I9 energy
(eV)
I9 FWHM
(meV)
Zn16
O-VS 3.35662 0.35
Zn16/18
O-VS 3.35903 0.48
Zn18
O-VS 3.35986 0.34
Zn16
O-VLS 3.35635 0.41
Zn16/18
O-VLS 3.35783 0.67
Zn18
O-VLS 3.35846 0.80
Table 5.6: Peak positions and FWHM of the I9 exciton recombination in the VS and
VLS O-enriched samples.
The first thing to note about these figures is that these previously reported
shifts are larger than those in our samples. This appears to indicate a lower level of
enrichment than expected, and additionally that the VLS samples have been enriched
to a lesser extent than the VS samples, compared to the enriched samples used in
previous reports. This is consistent with the reduced phonon frequency changes
observed in these samples in section 5.5, and is discussed further in chapter 6
regarding the levels of enrichment. The line widths of the BX lines in our VS and
134
VLS samples are, similarly to the Zn-enriched samples, very narrow (<1 meV) and
specifically much narrower than those reported by Manjón et al. (<5 meV) or Tsoi et
al. (2-8meV) in their single crystal samples. The very narrow line widths in our VS
and VLS samples once again demonstrate their excellent optical quality and their
potential for use in detailed optical studies of defects and impurities using low
temperature PL.
5.7 Conclusions
Two methods have been developed to ZnO nanorod samples isotopically
enriched with 18
O. Firstly the modified three-step process used to grow Zn-enriched
ZnO nanorods as discussed in chapter 4 has been further modified, by replacing the
atmospheric O2 with enriched 18
O2, in order to grow O-enriched ZnO nanorods using
this VS method on CBD buffer layers. In addition, O-enriched ZnO nanorods were
successfully grown in collaboration with our colleagues in Jena, using isotopically
enriched source powders, produced in DCU, in their VLS growth system. SEM
studies confirmed the success of both growth methods in terms of nanostructure
morphology, although in the case of the VLS samples, the nanorods were not c-axis
aligned due to the nature of the substrate used. A variety of XRD scans of the VS
samples indicated that these samples are not strained at the limit of detection of our
XRD system. miniSIMS spectra clearly showed the presence of 16
O in the natural
samples, and although it showed the presence of 18
O as expected in enriched
samples, it did not give direct confirmation of enrichment because of the continued
presence of 16
O contamination from a number of sources. However, Raman and PL
studies indicated clearly that O-enrichment was successful in both cases, although
the figures indicate the enrichment may be at a lower level in our samples compared
to previous reports with the same nominal enrichment levels. This is discussed
further in the next chapter. The very narrow line widths in our VS and VLS
demonstrate their excellent optical quality and there potential for use in detailed
optical studies of defects and impurities using low temperature PL.
135
5.8 References
1 T.-W. Koo and D. Whang, Nano 03, 477 (2008).
2 A. Umar, S.H. Kim, and Y.B. Hahn, Superlattices Microstruct. 39, 145 (2006).
3 R.B. Saunders, S. Garry, D. Byrne, M.O. Henry, and E. McGlynn, Cryst. Growth
Des. 12, 5972 (2012).
4 C. Borchers, S. Mu, D. Stichtenoth, D. Schwen, and C. Ronning, J. Phys. Chem. B
110, 1656 (2006).
5 R.B. Saunders, Theoretical and Experimental Studies of ZnO Nanowires Grown by
Vapour Phase Transport, Dublin City University, 2012.
6 M.A. Zimmler, T. Voss, C. Ronning, and F. Capasso, Appl. Phys. Lett. 94, 241120
(2009).
7 B.-H. Hwang, J. Phys. D. Appl. Phys. 34, 2469 (2001).
8 R.T.R. Kumar, E. McGlynn, M. Biswas, R. Saunders, G. Trolliard, B. Soulestin, J.-
R. Duclere, J.P. Mosnier, and M.O. Henry, J. Appl. Phys. 104, 084309 (2008).
9 E. McCarthy, R.T. Rajendra Kumar, B. Doggett, S. Chakrabarti, R.J. O’Haire, S.B.
Newcomb, J.-P. Mosnier, M.O. Henry, and E. McGlynn, J. Phys. D. Appl. Phys. 44,
375401 (2011).
10 International Centre for Diffraction Data. Joint Commitee on Powder Diffraction
Standards (JCPDS) Database. Available at: http://www.icdd.com/
11 V. Mote, Y. Purushotham, and B. Dole, J. Theor. Appl. Phys. 6, 6 (2012).
12 S. Desgreniers, Phys. Rev. B 58, 14102 (1998).
13 B.D. Cullity and S.R. Stock, Elements of X-Ray Diffraction, 3rd ed. (Prentice Hall,
New Jersey, 2001).
14 J. Serrano, F. Widulle, A.H. Romero, A. Rubio, R. Lauck, and M. Cardona, Phys.
Status Solidi 235, 260 (2003).
15 J. Serrano, F. Manjón, A. Romero, F. Widulle, R. Lauck, and M. Cardona, Phys.
Rev. Lett. 90, 055510 (2003).
16 G. Xing, G. Xing, M. Li, E.J. Sie, D. Wang, A. Sulistio, Q. Ye, C.H.A. Huan, T.
136
Wu, and T.C. Sum, Appl. Phys. Lett. 98, 102105 (2011).
17 N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, and G. Cantwell,
Appl. Phys. Lett. 81, 622 (2002).
18 S. Muller, D. Stichtenoth, M. Uhrmacher, H. Hofsass, C. Ronning, and J. Roder,
Appl. Phys. Lett. 90, 012107 (2007).
19 M. Schilling, R. Helbig, and G. Pensl, J. Lumin. 33, 201 (1985).
20 M. Biswas, Y.S. Jung, H.K. Kim, K. Kumar, G.J. Hughes, S. Newcomb, M.O.
Henry, and E. McGlynn, Phys. Rev. B 83, 235320 (2011).
21 J. Cullen, D. Byrne, K. Johnston, E. McGlynn, and M.O. Henry, Appl. Phys. Lett.
102, 192110 (2013).
22 B.K. Meyer, J. Sann, S. Lautenschläger, M.R. Wagner, and A. Hoffmann, Phys.
Rev. B 76, 184120 (2007).
23 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
Cardona, Solid State Commun. 128, 35 (2003).
24 S. Tsoi, X. Lu, A.K. Ramdas, H. Alawadhi, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 74, 165203 (2006).
137
Chapter 6: Optical Properties and
the Cu-related Defect in Isotopically
Enriched ZnO
6.1 Introduction
The optical properties of isotopically enriched ZnO nanorods are a central
focus of this work, as the high optical quality of the material produced allows
detailed study of such properties at low temperatures. This chapter presents the full
set of low temperature PL data, from both the SPEX and FT spectroscopic systems,
and from the Zn- and O- enriched samples grown by the three novel methods as
presented in chapter 4 and 5. Specifically, the near band edge data is first presented,
followed by the SGB ZPL, data for each set. These results are used in subsequent
study of the Cu-related defect at 2.86 eV regarding any involvement of native
defects. Not every spectrum for every sample is presented, but a sub-set
representative of those observed in all samples is shown. Both the SPEX and FT
systems were used to record spectra of the band edge regions and the green band
ZPLs of each sample and the spectra are treated as described in chapter 2.
There has been a reasonable degree of prior study of FX and Raman feature
behaviour in isotopically enriched ZnO, but much less work done on BX emissions.
138
Based on previous work in the field1–3
, we have worked on the assumption that the
FX binding energy is unaffected by isotopic mass and thus that the FX and shallow
BX features follow the band gap variation with isotopic mass. Such an assumption is
not generally valid for all materials4–6
but has been shown to be valid for ZnO and
allows us to compare our BX and deep emission data with the FX behaviour reported
previously.
6.2 Zn-enriched Nanorods
6.2.1 PL Spectra
Figure 6.1 shows the band edge and Cu-related 2.86 eV ZPLs as recorded on
the SPEX and FT systems for the Zn-enriched samples. (Figure 6.1(a) is the same
as previously presented in figure 4.15.) Figure 6.1(a) and 6.1(b) show representative
spectra from the SPEX and FT systems from the band edge region using the ‘pure’
isotope samples. The samples with mixed isotopic distributions produced similar
spectra. The SPEX system produces much less noisy spectra than the FT. The
dominant feature in the spectra is the I9 line attributed to In impurities which is of
particular interest, as shown below.7 The I8 and I6 lines attributed to Ga and Al
impurities are also clearly observed.7–9
The I2 line attributed to ionised In
impurities10
and the surface exciton11
(SX) emission are also visible and labelled in
the figure. These impurities most probably come from the laboratory environment
where the growth takes place, for example from the alumina boat, or the source
powders. Overall the SPEX spectra are very detailed and have good signal to noise
ratios, while the FT spectra suffer from slightly more noise, probably due to the
smaller dynamic range of the instrument and the very high intensity of the band edge
emissions from these samples, which is particularly apparent when shown on a log
scale as in this figure. However, the position of the I9 line can still be determined
with high accuracy. The I6 line is the only other readily identifiable line in the FT
spectra.
139
Figure 6.1: Band edge (a and b) and Cu-related 2.86 eV ZPL (c and d) spectra from
the Zn-enriched samples from the SPEX system (a and c) and the FT system (b and
d), (spectra shifted vertically for clarity).
In contrast to the band edge spectra, the FT system is more suited to
recording the deep level ZPL spectra, as shown in figure 6.1(c) and (d). This stems
mainly from the diffraction grating in the SPEX monochromator, which is blazed so
to be most effective in the band edge region. It is much less efficient around the ZPL
and green band region. It was therefore difficult to obtain spectra of the ZPL on the
SPEX system. A portion of each sample was annealed for ten minutes at 900˚C to
try and increase the green band intensity for the SPEX scans.12
This was successful
in the cases of 64
ZnO, 66
ZnO and 68
ZnO. The ZPLs shown for these samples have
been taken from annealed samples. The ZPLs for the other samples are not from
annealed samples. This is because, in those cases, the samples gave more intense,
less noisy spectra before annealing than after. This could be due to the self-limiting
nature of the SGB intensity in ZnO.13
As the concentration of Cu grows, the
140
luminescence intensity increases but then begins to decrease. The concentration of
Cu was not controlled in this experiment. Nevertheless, the intensities of the ZPLs
in the mixed samples are notably weaker than the others, and their spectra are
therefore noisier. Variation in intensity between samples is also observed in the FT
spectra, where none of the samples were annealed. However the FT spectra of the
ZPLs were much less noisy than the SPEX ones for most samples, such as the mixed
isotope samples shown here for example. Overall the variations in intensity could be
attributed to differences in Cu content and slight differences in optical alignment
between samples.
6.2.2 Discussion on Energy Shifts and Involvement of Zni or VO in
the Cu-related Emission at 2.86 eV
In order to determine if there is any involvement of Zni or VO in the Cu-
related 2.86 eV ZPL and associated SGB in ZnO, we compare the shifts and line
widths in exciton recombination energies in the band edge region with those of the
2.86 eV ZPL. Band edge energy shifts were observed in these samples (both Zn-
enriched and O-enriched) when the isotopic content of the crystal was changed, as
presented in chapters 4 and 5, due to the change in mass and corresponding change
in the energy of the vibrational states involving the defect and the surrounding lattice
due to the isotope effect as discussed in chapter 3. This isotope effects both the BX
positions in the band edge, as well as the Cu-related ZPL. However if a native
defect, such as an interstitial or vacancy, was also involved in the 2.86 eV ZPL, in
addition to the Cu substitutional atom, then we could expect a different shift and also
possibly a different line width to occur here. This is because the additional native
defect would result in a change in the local configuration and/or vibrational
properties at this defect site that would not be present for the shallow donor bound
defects. We can therefore look for any evidence of native defect involvement in the
2.86 eV ZPL by examining the positions and line widths of the BX at the band edge
and the 2.86 eV ZPL and looking for any relative energy shift between them or
change in line width when the isotopic content of the surrounding crystal is changed.
141
Any relative shifts could indicate the presence of Zni or VO native defects
complexing with the Cu atom in this defect. A substitutional Cu atom on a Zn site is
surrounded by O atoms in the lattice. An additional Zn atom complexing with the
Cu atom, or indeed a missing O atom (and the corresponding effects on the Zn atoms
surrounding the O vacancy), could affect the local vibrational environment in such a
way as to produce an anomalous shift relative to the vibrational changes induced in
the background lattice by changes in the Zn isotopic masses in the lattice. Only
changes in the background lattice would affect the I9 BX, which is known with
confidence to be due to an isolated In atom substituting on a Zn site (InZn)14
, with no
complexing with native defects, and in any case this bound exciton is loosely
localised and thus samples a large crystal area and is less sensitive to the local
environment of the InZn. Thus the shift in the I9 line serves as a natural reference
reflecting the changes in the background lattice only. However, an additional native
defect in the Cu defect would produce a distinct change in vibrational environment
since the carriers in this defect are much more strongly localised and thus will be
strongly affected by both the local defect environment as well as the changes in the
background lattice, and hence the relative energy shift between the defect ZPL and
the I9 emission provides a method whereby the presence of such local complexing
phenomena can be studied.
The position of the I9 In BX recombination was selected as a reference
energy to measure changes in the band edge positions with changing isotope
enrichment. The positions and line widths of the I9 line and the Cu-related 2.86 eV
ZPL were measured in each sample by fitting Gaussian curves to these peaks. This
was done for both the SPEX spectra and the FT spectra separately.
142
Table 6.1: Energies and widths of the I9 and ZPL lines in Zn-enriched ZnO, from the
SPEX system.
Table 6.2: Energies and widths of the I9 and ZPL lines in Zn-enriched ZnO, from the
FT system.
Sample
Average Zn
isotopic content
(amu)
I9
energy
(eV)
I9 FWHM
(meV)
ZPL energy
(eV)
ZPL
FWHM
(meV)
64ZnO 64 3.35629 0.44 2.85888 0.30
66ZnO 66 3.35662 0.36 2.8592 0.30
68ZnO 68 3.35689 0.41 2.85942 0.30
64/66ZnO 65 3.35641 0.33 2.85917 0.20
66/68ZnO 67 3.35682 0.33 2.85929 0.43
64/68ZnO 66 3.35643 0.35 2.85914 0.10
64/66/68ZnO
66 3.35603 0.35 2.85909 0.34
natZnO
65.4 3.35650 0.31 2.85897 0.31
Sample
Average Zn
isotopic content
(amu)
I9 energy
(eV)
I9 FWHM
(meV)
ZPL energy
(eV)
ZPL
FWHM
(meV)
64ZnO 64 3.35616 0.25 2.85897 0.60
66ZnO 66 3.35660 0.25 2.85925 0.47
68ZnO 68 3.35683 0.38 2.85951 0.53
64/66ZnO 65 3.35635 0.30 2.85913 0.51
66/68ZnO 67 3.35666 0.22 2.85940 0.47
64/68ZnO 66 3.35647 0.30 2.85922 0.50
64/66/68ZnO
66 3.35655 0.34 2.8592 0.55
natZnO
65.4 3.35657 0.33 2.85918 0.53
143
For the Zn-enriched samples, the positions and line widths of each of the I9
lines and ZPL lines, for both the SPEX and FT systems, are shown in tables 6.1
(SPEX) and 6.2 (FT). The positions of the lines from the SPEX and FT systems are
quite consistent. The average difference between the readings with the two systems
for the I9 line peaks is 0.13 meV, with the largest difference being 0.52 meV.
Similarly, for the ZPLs, the average is 0.1 meV with the largest difference being 0.21
meV. There is therefore very good agreement on the line positions between the two
systems and the line widths are also consistent across all samples for both BX and
ZPL emissions. We note that the optical quality of the material is excellent, with BX
line widths from the SPEX system given in section 4.6 are far smaller (< 1 meV)
than those reported by previous authors of < 5 meV, and 2-8 meV.15,16
No particular
trend is evident regarding I9 or ZPL line width changes with isotopic mass.
The following figures present the data from tables 6.1 and 6.2 in visual
format. Figure 6.2(a) contains data from the SPEX system and figure 6.2(b) contains
data from the FT system. In each graph, the left hand y-axis (black) shows the I9
energies, and the right hand y-axis (blue) shows the ZPL energies. The I9 axes both
cover the same range (3.3559 - 3.357 eV), and the ZPL axes also cover this range of
1.1 meV but with these axes shifted to align the ZPL energy and the I9 energy in the
64ZnO sample at the same point on each graph. In this way, it is easy to observe the
shifts of both the I9 and ZPL energies in each sample with changing isotopic
composition, and to compare any changes between them. The error bars of the
graphs represent the resolution for each scan to give an indication of the accuracy of
the positions. These are again ±0.035 meV for the I9 lines and ±0.045 meV for the
ZPLs on the SPEX and ±0.007 meV for the I9 lines and ±0.07 meV for the ZPLs on
the FT.
We first consider the results for the SPEX system in figure 6.2(a). The I9
energy increases with increasing Zn isotopic mass from 64
ZnO to 68
ZnO by 0.6 meV
(as described in section 4.6). Similarly, the ZPL energy increases by 0.54 meV. The
difference energy change here is 0.06 meV. Now turning to the FT system, in figure
6.2(b) the I9 energy increases when moving from 64
ZnO to 68
ZnO by 0.67 meV and
the ZPLs increase by 0.54 meV. The difference in energy change here is 0.13 meV.
144
Figure 6.2: Energies of the I9 lines (black) and Cu-related ZPL lines (blue) in the
Zn-enriched ZnO samples, from the (a) SPEX and (b) FT systems.
The trends shown in figure 6.2 clearly indicate that there is a consistent
increase in I9 and ZPL line positions from 64
ZnO to 68
ZnO of ~0.6 meV and that the
relative energy shifts are as small as ~0.1 meV. This latter figure is in the region of
the difference in line position between scans with the SPEX and FT systems. It is
also close to the step sizes of the ZPL scans on each system, indicating that it is close
to the experimental error considering the scan resolutions. These trends are clearly
shown when considering the 64
ZnO, 66
ZnO and 68
ZnO samples as marked. There are
some outliers, particularly with the 1/3 mixed isotope sample in figure 6.2(a). This
could be attributed to the very weak ZPLs recorded by the SPEX for these samples
making it difficult to determine the peaks clearly due to noise. It is also the case in
the mixed samples that, despite careful mixing of the powders, the amount of each
145
powder consumed during growth and therefore deposited on the sample may not
have been exactly equal. This may result in the average isotopic mass being slightly
different than desired, and hence the positions on the graphs.
The figures from both SPEX and FT data are consistent with each other. The
increases in band edge and ZPL positions are of the same order as those reported
previously for ZnO single crystals.15,16
There appears to be neither a significant shift
in ZPL position relative to the band edge position nor any change in line width of
this feature. Furthermore, there is no significant change in line width of the I9 line
across the set of samples. Taken together, the absence of a discernible relative
spectral shift, in addition to the constant line width of the I9 and ZPL features for all
samples, along with previous work showing that the Cu atom in this defect is on a Zn
site and therefore surrounded by O atoms, this result provides very strong evidence
that there is no involvement of Zni or VO native defect complexing in the Cu-related
2.86 eV ZPL and SGB emission in ZnO.
6.3 O-enriched Nanorods
6.3.1 PL Spectra
Figure 6.3 displays typical band edge and ZPL spectra as recorded on the
SPEX and FT systems for the O-enriched samples. Specifically, figures 6.3(a) and
6.3(b) show the band edge regions in the VS and VLS samples from the FT system
(the corresponding spectra from the SPEX data were presented in figure 5.12). As
previously, the BX region is dominated by the I9 In line and the I6 Al line.9,14
Some
other BX lines also appear in some samples, similarly to the spectra presented in
section 5.6. Figure 6.3(c) shows the ZPL line in the VS samples on the SPEX
system and figure 6.3(d) shows those in the VLS samples on the FT system. These
are typical of the ZPL lines seen in all the O-enriched samples on both systems.
Note that the other lines in the FT spectra in this region are from the Hg calibration
lamp. The FT band edge spectra from these samples are less noisy than those from
the Zn-enriched samples and the BX lines are more clearly determined, perhaps due
146
to stronger band edge emission in these samples. The VS samples were all annealed
to activate the SGB as described in section 5.6, while the VLS samples required no
such anneal as the SGB was present as grown. Similarly to the Zn-enriched samples,
these BX and Cu impurities are common in ZnO and most probably are due to the
growth environment, i.e. source powders and the boats and tubes used.
Figure 6.3: Band edge (a and b) and Cu-related 2.86 eV ZPL (c and d) spectra from
the VS and VLS O-enriched samples from the FT system (a, b and d) and the SPEX
system (c). (Spectra shifted vertically for clarity. Artefact in mixed sample in (a) is
due to FT apodisation. Vertical lines in (c) are from Hg calibration lamp.)
Overall, in figure 6.3, the high optical quality of the samples is clearly
apparent, and the shifts to higher energies with increasing O isotopic mass, for both
the BX and Cu ZPL regions, are readily apparent. We note that the optical quality of
the material is excellent, with BX line widths from the SPEX system given in section
5.6 are far smaller (< 1 meV) than those reported by previous authors of < 5 meV,
147
and 2-8 meV.15,16
Also of note is the clear increase in the line width of the ZPL lines
moving from the Zn16
O samples to the Zn18
O samples, which is not seen for the BX
emissions. This is discussed further below.
6.3.2 Energy Shifts
The peak positions and line widths of the I9 and Cu ZPL lines in both the VS
and VLS samples are shown in tables 6.3 and 6.4, for the SPEX and FT systems,
respectively. Once again, the peaks were fitted with Gaussians to determine the
peak positions and line widths as for the Zn-enriched samples. Similarly to the Zn-
enriched samples, there is good agreement between the SPEX and FT systems
regarding the peak energies for these lines. The average difference across each
sample between the two systems for the I9 lines was 0.2 meV, with the largest
difference being 0.31 meV. For the ZPLs, the average difference between the two
systems is 0.17 meV, with the largest being 0.36 meV. These figures are very
similar to those observed in the Zn-enriched samples and this demonstrates the good
and consistent agreement between the systems.
Table 6.3: Energies and widths of the I9 and ZPL lines in O-enriched ZnO, from the
SPEX system.
Sample
Average O
isotopic content
(amu)
I9
energy
(eV)
I9
FWHM
(meV)
ZPL
energy
(eV)
ZPL
FWHM
(meV)
Zn16
O-VS 16 3.35662 0.35 2.85915 0.44
Zn16/18
O-VS 17 3.35903 0.48 2.86070 0.79
Zn18
O-VS 18 3.35986 0.34 2.86162 1.18
Zn16
O-VLS 16 3.35635 0.41 2.85895 0.36
Zn16/18
O-VLS 17 3.35783 0.67 2.85993 0.76
Zn18
O-VLS 18 3.35846 0.80 2.86032 0.93
148
Table 6.4: Energies and widths of the I9 and ZPL lines in O-enriched ZnO, from the
FT system.
The set of graphs below show the energies of the I9 and Cu-related ZPL lines
in each O-enriched sample from tables 6.3 and 6.4. Figure 6.4 shows the data from
the (a) VS, and (b) VLS samples recorded on the SPEX system and figure 6.5 shows
the same data from the FT system. In each graph, the left hand y-axis (black) shows
the I9 energies, and the right hand y-axis (blue) shows the ZPL energies. The I9 axes
all cover the same range (3.356 - 3.3605 eV), and the ZPL axes also cover this range
of 4.5 meV but with these axes shifted to align the ZPL energy and the I9 energy in
the Zn16
O sample at the same point on each graph. Once again, it is easy to observe
the shifts of both the I9 and ZPL energies in each sample with changing isotopic
composition, and to compare any changes between them. As with the Zn-enriched
samples, the error bars of the graphs represent the step size for each scan to give an
indication of the accuracy of the positions. These are again ±0.035 meV for the I9
lines and ±0.045 meV for the ZPLs on the SPEX and ±0.007 meV for the I9 lines and
±0.07 meV for the ZPLs on the FT.
Sample
Average O
isotopic content
(amu)
I9
energy
(eV)
I9
FWHM
(meV)
ZPL
energy
(eV)
ZPL
FWHM
(meV)
Zn16
O-VS 16 3.35648 0.41 2.85879 0.68
Zn16/18
O-VS 17 3.35913 0.41 2.86057 1.08
Zn18
O-VS 18 3.36013 0.37 2.86140 1.14
Zn16
O-VLS 16 3.35649 0.95 2.85897 0.58
Zn16/18
O-VLS 17 3.35752 1.53 2.85979 1.03
Zn18
O-VLS 18 3.35820 1.31 2.86020 1.09
149
Figure 6.4: Energies of the I9 lines (black) and Cu-related ZPL lines (blue) in the (a)
VS, and (b) VLS, O-enriched ZnO samples, from the SPEX system.
The increases in the energies of both the I9 and Cu-related ZPL lines are
clearly shown in figures 6.4 and 6.5. This indicates that isotopic enrichment with
different compositions of O isotopes has been successful. However, there are a
number of things to note while examining these energy shifts as described in the next
section. Table 6.5 summarises the shifts observed between the Zn16
O and Zn18
O
samples, for the I9 and ZPL lines, in both the VS and VLS samples, and with both
the SPEX and FT systems. The ratio of the I9 shifts to the ZPL shifts in each case is
also shown in the table.
150
Figure 6.5: Energies of the I9 lines (black) and Cu-related ZPL lines (blue) in the (a)
VS, and (b) VLS, O-enriched ZnO samples, from the FT system.
I9 energy shift from
Zn16
O to Zn18
O (meV)
ZPL energy shift from
Zn16
O to Zn18
O (meV)
Ratio
I9/ZPL
SPEX
VS samples 3.24 2.47 1.31
VLS samples 2.11 1.37 1.54
FT
VS samples 3.65 2.61 1.40
VLS samples 1.71 1.23 1.39
Table 6.5: Summary of the band edge and ZPL energy shifts in the O-enriched
samples.
151
Finally, we note again that the ZPLs for the O-enriched samples, presented in
figure 6.3(c) and (d), display a broadening effect moving from Zn16
O to Zn18
O,
which is not seen for the BX emissions. This suggests possibly poorer enrichment in
the samples than expected, as the presence of both isotopes would produce this line
broadening effect, as discussed below (essentially because the ZPL lines are highly
susceptible to their local environment.), whereas no such broadening would be
expected if the 18
O sample were close to an actual 100% enrichment. For example,
in figure 6.3(c) (SPEX, VS samples) the ZPLs broaden from 0.24 meV to 1.25 meV
moving from Zn16
O to Zn18
O, and in figure 6.3(d) (FT, VLS samples) the ZPLs
broaden from 0.58 meV to 1.09 meV moving from Zn16
O to Zn18
O. These figures
are typical of those for these samples with both spectrometers as shown in tables 6.3
and 6.4. No particular trend is evident from the I9 line widths.
6.3.3 Discussion on Energy Shifts, Line Widths and Involvement of
Oi or VZn in the Cu-related Emission at 2.86 eV
The first thing to note here is that the I9 energy shifts between the Zn16
O and
Zn18
O samples are less than those previously reported in the literature for the same
nominal degree of 18
O enrichment. For example, Manjón et al. reported BX shifts
over this range of ~6.4 meV using the I4 recombination, and Tsoi et al. reported a
figure of ~6.5 meV for shifts in the A-exciton energy.15,16
However, in our VS
samples, the I9 energy shifts are recorded as 3.24 meV in the VS samples, and just
2.11 meV in the VLS samples using the SPEX data moving from 16
O to 18
O. The
corresponding figures from the FT system are listed in table 6.5 and follow this
general trend. This would appear to indicate that our samples are not as fully
enriched as expected from the growth processes as designed and carried out. This is
consistent with the E2high
frequencies presented in chapter 5 which were also less
than reported previously. The Zn16/18
O energies also appear closer to Zn18
O than
Zn16
O, i.e. the shift with isotopic mass is nonlinear. This is also observed by Manjón
et al.
Secondly, the VLS samples grown in Jena consistently display a smaller shift
than the corresponding one in the VS samples grown in DCU. This is observed in
152
both the SPEX and FT spectroscopic systems, and both for the I9 and SGB ZPL
recombinations. This suggests that not only are both the VS and VLS samples
enriched to a lesser degree than thought, but that within this work, the VLS samples
are themselves enriched to a lesser degree than the VS samples. The phonon
energies measured by Raman spectroscopy for these samples presented in chapter 5
are also consistent with this.
Thirdly, it is clear that the shifts in the ZPL energies in all the O-enriched
samples are less than the corresponding shifts in the band edge region. This is
observed across the VS and VLS samples, and on both spectroscopic systems. For
example, using the SPEX data, the I9 shift in the VS samples is 3.24 meV, but the
ZPL shift is just 2.47 meV. Likewise, the I9 shift in the VLS samples is 2.11 meV,
but the ZPL shift is just 1.37 meV. The ratio of the I9 shifts to the ZPL shifts in table
6.5 shows a consistent relative shift of between the two in each case, with a slightly
larger relative shift recorded on the SPEX for the VLS samples. There are also clear
increases in the ZPL line widths, which are not seen in the BX emissions. There are
various ways to interpret these data. The first is that this indicates the presence of Oi
or VZn native defects in complex with the Cu atom at this defect. Since the CuZn
defect is a substitutional Cu atom on a Zn site, it is surrounded by O atoms in the
lattice. An additional O atom, or a missing Zn atom (and the corresponding effects
on the O atoms surrounding the Zn vacancy), could affect the local vibrational
environment in such a way as to produce an anomalous shift and line broadening. In
principle each possible configuration of the surrounding O atoms would lead to a
distinct line17
, allowing a study of the possible numbers of O atoms involved, but the
small shifts associated with different configurations mean that these overlap and
simply broaden the ZPL emission. A second possibility is that there exists a number
of different configurations of the 4 O atoms surrounding the Cu due to the varying
possible combinations of 16
O and 18
O atoms that the Cu atom could be adjacent to
(e.g. 4 x 16
O, 1 x 18
O + 3 x 16
O etc.). The effect of the varying O-isotopic masses in
the 4 closest O atoms in the vicinity of the Cu defect could be the cause of the
anomalous shifts and the line broadening. Once again in principle each possible
configuration of the surrounding atoms would lead to a distinct line,17
allowing a
study of the possible numbers of O atoms involved, but the small shifts associated
with different configurations mean that these overlap and simply broaden the ZPL
153
emission. Line broadening of this type is expected for a deep defect with localised
carriers strongly affected by the local defect configuration, whereas it is not expected
for the shallow BX emissions where the small exciton confinement energy means
that the local defect configuration is less important, and the BX samples a larger
volume of the crystal and the line energy is affected by the average isotopic mass in
this volume, but the line width is unaffected.
Ultimately, because of the small line shifts and the inability to resolve lines
associated with individual O atomic configurations we do not have enough
information to determine which of these possibilities is the case, although we
consider it more likely that it is the latter as it would agree with the previous work of
Dingle18
and Byrne et al.19
, in terms of the defect symmetry (C3v)20
and other known
aspects of the defect. The likelihood of a complex of a Cu atom with 3 native defects
(Oi or VZn) producing a defect with overall C3v symmetry seems very small. By
contrast, the starting configuration of the 4 O atoms surrounding the Zn site on
which the Cu atom substitutes has this symmetry naturally, and it seems a far more
likely explanation that any relaxation of the surrounding O after Cu substitution will
maintain this symmetry.
Finally, as discussed above, because the ZPL for the nominally 100% enriched
18O sample shows a considerable broadening (whereas no such broadening would be
expected if the 18
O sample were close to an actual 100% enrichment) we believe that
the samples are not as highly enriched as the nominal enrichment values, and we
estimate the actual 18
O enrichments below.
6.3.4 Estimate of Level of Enrichment in O-enriched ZnO
Several aspects of our data appear to point towards the nominally 100% 18
O-
enriched ZnO nanorods not being 100% enriched with the desired isotopic
composition. These data include Raman and band edge PL shifts being smaller than
those reported previously in the literature, and the broadening of the Raman and ZPL
lines with increased 18
O enrichment. It is useful to make an estimate of the level of
enrichment achieved in our samples by comparing our band edge region BX energy
154
shifts to those reported by Manjón et al. and Tsoi et al, assuming these data reported
previously are from samples actually enriched to nominally 100%.15,16
Moving from
Zn16
O to Zn18
O, the average shift in those two reports is ~ 6.45 meV. Using this as a
reference, table 6.6 then gives the shifts observed in the band edge region in our VS
and VLS samples as a percentage of this, for both the Zn18
O and the Zn16/18
O mixed
isotope samples, from both spectroscopic systems The same calculation can also be
carried out using the E2high
phonon energy shifts in section 5.5 compared to the room
temperature value of ~23 cm-1
reported by Serrano et al.21
Across the two PL
spectroscopic systems, and the Raman data, this gives an average estimated
enrichment of 60% for Zn18
O-VS, 32% for Zn18
O-VLS, 41% for Zn16/18
O-VS and
21% for Zn16/18
O-VLS. The ratios of the 18
O enrichment in the VS samples to the
VLS samples in each case is also shown in table 6.6, and range from 1.5-2.1 in the
18O samples and 1.6-2.6 in the
16/18O samples. The average ratios are 1.9 in the
18O
samples and 2.1 in the 16/18
O samples.
18O-enrichment as % of
literature reference for
Zn18
O nanorods
18O-enrichment as % of
literature reference for
Zn16/18
O nanorods
SPEX PL
VS samples 50 37
VLS samples 33 23
Ratio VS/VLS 1.5 1.6
FT PL
VS samples 57 41
VLS samples 27 16
Ratio VS/VLS 2.1 2.6
Raman
VS samples 72 46
VLS samples 37 23
Ratio VS/VLS 2.0 2.0
Table 6.6: 18
O enrichment as % of literature PL and Raman references for Zn18
O
and Zn16/18
O nanorods.
155
6.4 Discussion on Possible Causes of Reduced Oxygen
Enrichment
As discussed above, a number of pieces of evidence presented in this chapter,
and in chapter 5, regarding O-enriched ZnO nanorods, point to the likelihood that
these samples are not enriched to the nominal levels stated, and expected from the
growth processes as performed. It is therefore worthwhile to discuss the possible
sources of this reduced 18
O enrichment. During both the growth of the VS samples
by method 1, and the oxidation of the Zn metal powders to produce the ZnO
powders for the VLS growths by method 2, as described in chapter 2, every effort
was made to ensure that only the appropriate O isotope was present.
Considering firstly method 1, the growth of 18
O-enriched ZnO nanorods by
the modified VS method, involving evacuating the furnace tube and refilling with
18O2, we must consider any possible sources of natural
16O2 as a contaminant which
could lead to reduced 18
O enrichment as observed in these samples. One obvious
source of this is the natural 16
O2 in the tube at the start of the experiment. The tube
was evacuated to a pressure of <1 mbar and then filled 21% atmosphere with 18
O2
and to atmosphere with N2. This would indicate a possible contamination level of
~0.1% 16
O2, or ~0.5% of the oxygen being this isotope. This certainly would not
explain the enrichment being reduced by about half as indicated by the PL data. The
N2 used in the experiment only has O2 impurities of <5 ppm, or around 0.0005%.
Clearly this would not be the source of such reduced enrichment in the atmosphere
inside the tube during growth. The system was extensively checked for leaks several
times during the set up and testing of the system. All line connections and seals
were thoroughly checked, and the lines and were purged of natural 16
O2 before
growth. This also included the regulator used on the 18
O2 cylinder. The natural 16
O
in the ZnO source powder (or adsorbed on the graphite powder) is also not likely to
be the cause of this effect, as the CTR reaction leads to the formation of CO
molecules, thereby trapping the O in the ZnO source powder and not allowing it to
react on the substrate, as the residual atmospheric O2 does. Attempts at growth with
just N2 gas in the tube, described in chapter 5, and the lack of observed growth from
such attempts, support this conclusion. We also considered whether residue on the
inside of the furnace tube (or some remaining in the alumina boat) from previous
156
growths could contribute some O2 to the growth by a small amount of it being
disassociated when heated, as the 18
O-enriched VS growths were carried out in the
same tube as the normal natural growths. However, the temperatures of 925°C were
not sufficient for this to occur. The final source of unwanted 16
O2 in the system
could simply be a leak. However, as mentioned above, the system was extensively
leak tested, and even if we consider, for example, 10 mbar of 16
O2 entering the
system during pump-down and refill, this would not come close to causing the
reduced PL and Raman shifts observed with these samples. Ultimately, we
emphasise once again that the fact that, when the system was tested using a pure N2
atmosphere with no O2 gas of any kind introduced, no VPT growth occurred,
strongly indicates that leaks or another unwanted sources of 16
O2 in the system were
not the cause of the poorer enrichment.
We now consider the VLS sample growths in Jena with our collaborators.
The system used to oxidise the Zn metal powder, using 16
O2 or 18
O2, to produce
Zn16
O and Zn18
O powders to be used in these growths was similar to the VS system
and there much of the previous paragraph also applies regarding leaks and the N2
gas. In this case, a smaller tube was used to reduce the amount of gas needed.
Additionally, a separate clean tube and boat was used for the 18
O isotope oxidisation,
to eliminate possible contamination. This, coupled with the lower temperatures of
~800°C, means we can be quite confident of having no contamination here. Initial
experiments on oxidising Zn powder were carried out using old powder which had
oxidised in air over time, however completely fresh Zn metal powder was used for
the 18
O oxidisation runs. It is considered likely that the cause of the even further
reduced enrichment in the VLS samples compared to the VS samples is that the
growth system in Jena led to increased 16
O contamination. These growths were
carried out using an alumina boat and tube that had deposits from previous growths
present. This is considered the most likely source of additional 16
O2 in the tube due
to the higher temperatures (~1350°C) needed for the sublimation of the ZnO sources
powders also producing some 16
O2 from the previous deposits.
Based on the discussion above, we are confident that our experimental
methods are sound and that we have effectively reduced or eliminated the obvious
sources of contamination, in terms of all the possible sources of 16
O2 while planning
the experiment, while the testing carried out before the growth proceeding indicated
157
that there was not a problem with 16
O2 contamination. If we consider the
explanation in the previous paragraph for the difference between the VS and VLS
samples, we must then consider the possibility that the cause of the reduced
enrichment in both the VS samples in DCU and the oxidised source powders
produced in DCU for the VLS growths in Jena originate from a common source.
Ultimately, it was considered that a possible source of 16
O2 was the lecture bottle of
18O2, and that perhaps this was not enriched properly to its rated value of >99%.
However we note that a point against this hypothesis is that the estimated enrichment
levels are not consistent with a fixed 16
O2 contamination of the 18
O2 gas used. This
would be a both a surprising and disappointing result, given that the gas bottle was
purchased from a reputable supplier and was clearly rated for a high enrichment
level and was properly handled at all stages.
Finally, between the submission of this thesis and the viva-voce examination,
a number of further tests were carried out in an attempt to clarify the source of the
16O contamination of these samples. Firstly, a further sample of Zn
18O-VS was
grown using the same CTR-VPT method except with a fresh bottle of 18
O2 gas
sourced from a different supplier (CK Isotopes, 98%). PL using the SPEX system
showed a band edge shift of 3.12 meV for the I9 line and 1.97 meV for the ZPL.
These figures are similar to those reported in this chapter and this test indicates that
the original bottle of 18
O2 gas was not contaminated with 16
O2. Secondly, the same
growth was repeated using a fresh tube and alumina boat which has no deposits from
previous unenriched growths. The results here were a band edge shift of 3.45 meV
for the I9 line and 1.79 meV for the ZPL which indicates that there is no
contamination from old deposits on the equipment. Finally a further VPT test was
carried out using a long 1 hour Ar flush at the start (versus the usual ~ 5 mins). This
removes all the residual oxygen from the tube and should result in no VPT nanorod
growth, and this is what was observed. This implies no other source of 16
O2 in the
system and is consistent with the nitrogen only test described in chapter 5 and
previous tests by others using this system.22
Ultimately, we must conclude that the
final source of the 16
O2 contamination was not conclusively identified during these
tests. We tentatively suggest that there could be some diffusion of 16
O from the
unenriched CBD buffer layer, or the Si native oxide layer in the VLS samples, which
could be contributing to this effect, especially since enhanced diffusion along the
158
nanorod sidewalls may be possible, compared to diffusion in single crystals.23
Alternatively some 16
O2 could be being released from these layers during the growth,
creating a concentration of 16
O2 in the immediate vicinity of the nanorods during
growth, although this possibility seems unlikely since it is not consistent with our
tests with either just N2 in the tube, or after a long Ar flush. Effects such as these
would be consistent with the thermal effects observed in the CBD layer when they
underwent a growth cycle which resulted in no VPT growth (i.e. with only N2 in the
tube or after a long Ar flush), where observable changes in morphology are seen, and
would also be consistent with the differing departures from the nominal enrichment
levels seen for the VS and VLS samples, since the efficiency of the processes might
well be different for the CBD buffer layer and the Si native oxide layer, used in these
two methods, respectively.
6.5 Conclusions
The low temperature PL data for the Zn-enriched, and both the VS and VLS
O-enriched ZnO nanorods were presented. The optical quality of all samples was
excellent, in that they each displayed intense band edge emissions with narrow BX
emission lines, far narrower than those reported previously in the literature. The
SGB was also present in all samples, although some samples required annealing to
increase its intensity to allow spectroscopy study. The Zn-enriched samples
displayed a shift of ~0.6 meV as the isotopic composition changes 64
ZnO to 68
ZnO,
in line with previously published work, for both the I9 BX recombination and the
SGB ZPL and no changes in line width of either emission with changes in isotopic
composition. This indicated no involvement of Zni or VO native defects in the Cu-
related defect at 2.86 eV. The O-enriched samples displayed band edge shifts of
around 3.5 meV as the isotopic content changed from Zn16
O to Zn18
O for the VS
samples, and smaller shifts of around ~2 meV for the VLS samples. These were
smaller than the shifts of >6 meV reported previously. This, coupled with the
phonon frequency data presented in chapter 5, and line width broadening in the ZPL
lines, indicated that the 18
O enrichment of these samples was not as expected. The
levels of such enrichment were estimated at of 60% for Zn18
O-VS, 32% for Zn18
O-
159
VLS, 41% for Zn16/18
-VS and 21% for Zn16/18
O-VLS. The reasons for this reduced
enrichment were discussed and it is concluded that the 18
O2 gas source used was the
most likely source, in addition to residue on the tube and boat used for the VLS
samples causing a further reduction in the enrichment of those samples. Smaller
relative shifts and line width broadening for the ZPL versus the band edge in the O-
enriched samples are attributed to changes in the local defect environment due to
different occupancy of O sites by 16
O and 18
O atoms in the mixed composition
samples, which seems a much more likely cause of the effects seen than complexing
with Oi or VZn defects, taking account of the known C3v defect symmetry.
6.6 References
1 F.I. Kreingol’d, Fiz. Tvend. Tela 20, 3138 (1978).
2 F.I. Kreingol’d and B.S. Kulinkin, Fiz. Tvend. Tela 28, 3164 (1986).
3 F.I. Kreingol’d, Fiz. Tvend. Tela 27, 2839 (1985).
4 F.I. Kreingol’d, K.F. Lider, and M.B. Shabaeva, in Fiz. Tvend. Tela (Cambridge,
1984), pp. 3490–3491.
5 A. Klochikhin and V.G. Plekhanov, Fiz. Tvend. Tela 22, 585 (1980).
6 V.G. Plekhanov, V.A. Pustovarov, A.A. O’Connell-Bronin, T.A. Betenekova, and
S.O. Cholakh, Fiz. Tvend. Tela 18, 2438 (1976).
7 B.K. Meyer, H. Alves, D.M. Hofmann, W. Kriegseis, D. Forster, F. Bertram, J.
Christen, A. Hoffmann, M. Straßburg, M. Dworzak, U. Haboeck, and A. V. Rodina,
Phys. Status Solidi 241, 231 (2004).
8 K. Johnston, M.O. Henry, D. McCabe, E. McGlynn, M. Dietrich, E. Alves, and M.
Xia, Phys. Rev. B 73, 165212 (2006).
9 M. Schilling, R. Helbig, and G. Pensl, J. Lumin. 33, 201 (1985).
10 J. Cullen, D. Byrne, K. Johnston, E. McGlynn, and M.O. Henry, Appl. Phys. Lett.
102, 192110 (2013).
160
11 M. Biswas, Y.S. Jung, H.K. Kim, K. Kumar, G.J. Hughes, S. Newcomb, M.O.
Henry, and E. McGlynn, Phys. Rev. B 83, 235320 (2011).
12 N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, and G. Cantwell,
Appl. Phys. Lett. 81, 622 (2002).
13 G. Xing, G. Xing, M. Li, E.J. Sie, D. Wang, A. Sulistio, Q. Ye, C.H.A. Huan, T.
Wu, and T.C. Sum, Appl. Phys. Lett. 98, 102105 (2011).
14 S. Muller, D. Stichtenoth, M. Uhrmacher, H. Hofsass, C. Ronning, and J. Roder,
Appl. Phys. Lett. 90, 012107 (2007).
15 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
Cardona, Solid State Commun. 128, 35 (2003).
16 S. Tsoi, X. Lu, A.K. Ramdas, H. Alawadhi, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 74, 165203 (2006).
17 S. Feofilov, A. Kulinkin, K. Ovanesyan, A. Petrosyan, and C. Dujardin, Phys.
Chem. Chem. Phys. 16, 22583 (2014).
18 R. Dingle, Phys. Rev. Lett. 23, 579 (1969).
19 D. Byrne, F. Herklotz, M.O. Henry, and E. McGlynn, J. Physics. Condens. Matter
24, 215802 (2012).
20 C. Solbrig, Z. Phys. A. 211, 429 (1968).
21 J. Serrano, F. Manjón, A. Romero, F. Widulle, R. Lauck, and M. Cardona, Phys.
Rev. Lett. 90, 055510 (2003).
22 R.B. Saunders, Theoretical and Experimental Studies of ZnO Nanowires Grown
by Vapour Phase Transport, Dublin City University, 2012.
23 A.C.S. Sabioni, M.J.F. Ramos, and W.B. Ferraz, Mater. Res. 6, 173 (2003).
161
Chapter 7: Conclusions and Future
Work
7.1 Conclusions
In this work, the three-step growth process previously developed in our
group1, itself based upon previous developments in the literature, has been
successfully adapted in order to grow ZnO nanorods containing different isotopes of
Zn and O by several novel methods. In doing so we have demonstrated the
possibility of using ZnO nanorods for studies of the effects of isotopic enrichment,
using relatively simple growth techniques, and have shown that the nanorods possess
excellent crystalline and optical quality, enabling detailed optical studies of both
band edge and deeper level emissions. A key original and novel feature of this
present work is that this is the first time isotopically enriched ZnO has been
produced in nanostructured form and we further note that the optical quality of these
nanorods, in terms of their band edge emission intensities, and more specifically
their line widths, is significantly better than that reported previously for ZnO bulk
single crystals.2,3
For the Zn-enriched nanorods, the growth process consisted of depositing a
ZnO seed layer by drop coating, followed by growth of a ZnO nanorod buffer layer
162
using CBD and the growth of ZnO nanorods by CTR-VPT. The previous growth
process was altered by reducing the amount of ZnO source powders used. Samples
of nat
ZnO, 64
ZnO, 66
ZnO, 68
ZnO, 64/66
ZnO, 66/68
ZnO, 64/68
ZnO and 64/66/68
ZnO were
grown, by altering the Zn isotopic composition of the ZnO sources powder, using
this VS method. The samples were characterised using SEM, XRD, SIMS, Raman,
low temperature PL and reflectance spectroscopy. SEM confirmed the dense
coverage of vertical, c-axis aligned nanorods over a large sample area. The
morphology was typical of that observed in previous work. Some variations in
morphology were observed such as differing heights of nanorods, nanorods
becoming ‘entangled’ and such loss of vertical alignment, reduced density and poor
nanorod morphology in one sample. However, these differences were mainly
confined to small areas of the samples and some sample to sample variation of this
nature is expected with our CTR-VPT growth process. The growth process is easy,
quick and very reliably produces high quality ZnO nanorods. XRD confirmed the
presence of wurtzite ZnO with an intense ZnO peak at 34.4˚. The Si peak at 69.1˚
was also clearly visible. The FWHM of the 34.4˚ ZnO peak rocking curves were
very narrow at ~2-3˚ and this confirmed the high levels of alignment and crystal
quality in the samples. SIMS data confirmed the successful Zn isotopic enrichment
consistently using two independent SIMS systems.
Raman data show a shift of >1 cm-1
in the peak position of the Raman
scattered peaks due to the E2low
and E2high
phonon modes when the Zn isotope is
changed from 64
Zn to 68
Zn, consistent with previous work on samples with different
isotopic enrichments, again confirming successful isotopic substitution.4,5
Low
temperature PL measurements, using both the SPEX monochromator and FT
spectrometer revealed the excellent optical quality of the samples, with line widths
of <1 meV. The results from the two systems were consistent and revealed increases
in the I9 and ZPL energies of ~0.6 meV with Zn isotopic content changing from
64ZnO to
68ZnO. This is similar to previous results for single crystals, although we
note previously reported figures of >1 meV.2,3
Despite this, we are confident of
good enrichment in these samples as supported by the SIMS and Raman data.
Relative to the band edge, the ZPL line shift reduces to ~0.1 meV in both systems.
We conclude that there is no detectable relative shift between the band edge and Cu-
related 2.86 eV ZPL when the Zn isotopic content of the lattice is changed. This
163
result allows us to conclude that we detect no involvement of the native defects Zni
or VO in this Cu-related defect.
Two novel methods were developed to produce O-isotopically enriched ZnO
nanorods. Following extensive development and testing, a further modified VS VPT
method was successful in growing ZnO nanorods enriched with 18
O. Samples of
Zn16
O, Zn16/18
O and Zn18
O were grown by evacuating the tube in the growth furnace
and refilling it with a mixture of ‘artificial’ atmosphere containing ~21% 18
O2 and
~79% N2 (or a 50:50 mix with natural O2). This made use of the fact that it is the
residual atmospheric O2 in the tube which contributes to the ZnO nanorods. A
second method of producing O-enriched ZnO nanorods was also developed in
collaboration with our colleagues in Jena, Germany. This was a VLS VPT method
based on the direct sublimation of ZnO source powder at higher temperatures. For
this, O-enriched ZnO powders were produced by oxidising Zn metal powder in DCU
prior to these growths taking place in Jena.
These O-enriched samples were again characterised using SEM, XRD,
miniSIMS, Raman spectroscopy and low temperature PL. SEM studies confirmed
the success of both growth methods in terms of nanostructure morphology, although
in the case of the VLS samples, the nanorods were not c-axis aligned due to the
nature of the substrate used. XRD again confirmed the growth of high quality
wurtzite ZnO nanorods. XRD scans of the VS samples indicated that these samples
are not strained at the limit of detection of our XRD system. miniSIMS spectra
clearly showed the presence of 16
O in the natural samples, and although it showed
the presence of 18
O as expected in enriched samples, it did not give direct
confirmation of enrichment because of the continued presence of 16
O contamination
from a number of sources. However, Raman and PL studies indicated clearly that O-
enrichment was successful in both cases, although the figures indicate the
enrichment may be at a lower level in the VLS samples compared to the VS samples,
as these displayed smaller shifts, for example 3.24 meV vs 2.11 meV in the PL band
edge region, from the SPEX monochromator, moving from Zn16
O to Zn18
O. The
very narrow I9 and other BX line widths in our VS and VLS again demonstrate their
excellent optical quality.
164
A number of factors suggest that the O-enrichment of the VS and VLS ZnO
nanorods may not be as great as expected. The band edge shifts above are notably
less than those previously reported of >6 meV for bulk single crystals.2,3
The
phonon energies in the Raman data are also less than those reported previously for
single crystals, and the line widths wider for samples containing 18
O.4,5
In addition,
the Cu-related ZPL at 2.86 eV displayed a line broadening effect in samples
nominally containing only 18
O. This also indicates the continuing presence of 16
O in
larger amounts than expected. Using the band edge shifts and Raman phonon energy
shifts in comparison to the literature, the level of 18
O enrichment in each sample was
estimated at 60% for Zn18
O-VS, 32% for Zn18
O-VLS, 41% for Zn16/18
-VS and 21%
for Zn16/18
O-VLS. The Cu-related ZPL shifts are also smaller than those at the band
edge in these O-enriched samples in addition to the line broadening in the mixed
isotope samples mentioned above. While these data could possible indicate Oi or
VZn defects at the defect site, we consider it more likely, based on considerations
related to the defect symmetry, that the effect of the varying occupancy of the 4
closest O atoms in the vicinity of the Cu defect by different O-isotopic masses is the
cause of the anomalous shifts and the line broadening.
Finally, a number of other experiments were carried out using samples
produced during this work and are presented in appendix A as examples of the
potential uses of ZnO nanorods grown using these methods in other studies. These
experiments were often carried out by collaborators using the samples produced
during this work. This included TRPL measurements carried out on the 64
Zno,
66ZnO and
68ZnO samples, CBD-grown ZnO nanorods produced used in experiments
on third harmonic UV generation, and some experiments carried out on particle
acceleration and the generation of x-rays from nanostructured targets using ZnO
nanorods grown using CBD on Ti targets.
7.2 Future Work
There are a number of further experiments proposed in this section in order to
provide a fuller understanding of the outcomes of the work contained in this thesis.
From our perspective the reduced O-enrichment in both the VS and VLS samples is
165
the most confusing result from this work and thus is one aspect that we would
consider a high priority to investigate further. We feel we have carefully considered
all the possible sources of natural 16
O enrichment in the VS growth apparatus, and
have tentatively suggested that there could be some diffusion of 16
O from the
unenriched CBD buffer layer, or the Si native oxide layer in the VLS samples, which
could be contributing to this. Alternatively (although less likely we feel) some 16
O2
could be being released from these layers during the growth, creating a concentration
of 16
O2 in the immediate vicinity of the nanorods during growth. This could have
affected both the VS samples grown in DCU and the VLS samples grown in Jena,
albeit to different extents, consistent with our data. A more detailed study of these
possible effects is suggested.
In addition to this, and given that our miniSIMS measurements did not provide
a conclusive measurement of the O-enrichment in those samples, a direct
measurement of such enrichment is still desirable, and specifically to check if such a
measurement concurred with our estimates of the enrichment levels based on the PL
and Raman shifts. This would also be of relevance to efforts to conclusively identify
the source of 16
O2 contamination during growth. Efforts to obtain such data by
measuring the enrichment in single nanorods in collaboration with a group with
access to an atom probe system are currently underway.
While we are confident that our Zn-enriched nanorods are enriched to high
levels because of our SIMS and Raman data, the band edge PL shifts are slightly
lower than those previously reported, and this could be a sign of lower enrichment
than thought. It would be an interesting experiment to get a further, independent and
direct measurement of the Zn-enrichment in these samples to investigate this
irregularity, although the SIMS and Raman data appear to be strongly supportive of
good enrichment in these samples. Once again, efforts to obtain such data by
measuring the enrichment in single nanorods in collaboration with a group with
access to an atom probe system are currently underway.
Finally, while we have produced Zn-enriched and O-enriched ZnO nanorods in
this work, the full set could be completed by producing Zn-enriched nanorods with
enriched 18
O. Assuming that if any problem with 16
O2 contamination during growth
was remedied, this would allow a complete study of the Raman and PL shifts in
166
isotopically enriched ZnO nanorods to be carried out, using all the possible isotopic
combinations. A more detailed study of the Cu-related centre at 2.86 eV could also
be done. If, in the future, we were certain of proper O-enrichment, then the band
edge vs. Cu SGB ZPL energy shifts could be studied without the possible effects of
the presence of multiple O isotopes.
7.3 References
1 D. Byrne, E. McGlynn, K. Kumar, M. Biswas, M.O. Henry, and G. Hughes, Cryst.
Growth Des. 10, 2400 (2010).
2 F.J. Manjón, M. Mollar, M.A. Hernández-Fenollosa, B. Marı, R. Lauck, and M.
Cardona, Solid State Commun. 128, 35 (2003).
3 S. Tsoi, X. Lu, A.K. Ramdas, H. Alawadhi, M. Grimsditch, M. Cardona, and R.
Lauck, Phys. Rev. B 74, 165203 (2006).
4 J. Serrano, F. Manjón, A. Romero, F. Widulle, R. Lauck, and M. Cardona, Phys.
Rev. Lett. 90, 055510 (2003).
5 J. Serrano, F. Widulle, A.H. Romero, A. Rubio, R. Lauck, and M. Cardona, Phys.
Status Solidi 235, 260 (2003).
A1
Appendix A: Some Applications of
these Growth Methods and Materials
in other Experiments
A.1 Introduction
Nanostructures such as ZnO nanorods lend themselves to many types of
experiments and measurements which can reveal useful and interesting data. This
chapter briefly presents the results of a number of other experiments carried out
using samples produced by the author during this work and supplied to other groups.
These illustrate just a few examples of the potential uses of ZnO nanorods grown
using these methods in other studies, often in very diverse fields. These examples
include measurements on the Zn-enriched ZnO nanorods, and other studies carried
out using CBD-grown ZnO nanorods on a number of substrates. The actual
experiments were generally carried out by collaborators using the samples produced
during this work, rather than by the author, but they are included here since some
time and effort went into sample production and in some cases published outputs
have resulted (see for example references 1–3).
Specifically, time-resolved photoluminescence (TRPL) was carried out on the
64Zno,
66ZnO and
68ZnO samples grown during this work by our colleague Dr.
Joseph Cullen in Linköping University in Sweden. CBD grown ZnO nanorods
produced during this work were used in experiments on third harmonic UV
generation by Prof. Enda McGlynn and colleagues at the Max Born Institute for
Nonlinear Optics and Short-Pulse Spectroscopy in Berlin, Germany, including Dr.
Ruediger Grunwald. Finally, some experiments were carried out on the generation
of x-rays from nanostructured targets using ZnO nanorods grown using CBD on
titanium (Ti) targets. These experiments were carried out by collaborators Dr.
A2
Matthias Schnürer and Julia Braenzel, also in the Max Born Institute in Berlin,
during a research visit there by the author in 2013.
A.2 TRPL of Zn-enriched ZnO nanorods
TRPL was carried out on the pure Zn-enriched samples in order to study the
BX decay lifetimes. Data from various samples is presented in figure A.1. Figure
A.1(a) shows the TRPL decay of the I9 BX recombination in 64
ZnO, 66
ZnO and
68ZnO nanorods. Figure A.1(b) shows the same data for the first LO replica of the I9
line. The decay in all cases can be described by two exponentials, a fast component
τf, and a slow component τs.4 These time constants and their relative amplitudes (Aτf
and Aτs) for the I9 BX recombination in each sample are shown in table A.1. The
fast component can be attributed to surface effects and the slow component to the
intrinsic BX radiative recombination rate. The figures in table A.1 are consistent
with those previously reported for ZnO nanorods, and we note this report also
contains a study on the changes in lifetimes with nanorods size.4 The I9 exciton
displays an almost single exponential decay as shown by the (close to) straight lines
in figure A.1(a), which indicates that the decay is dominated by the slow component
indicating very high quality samples. The 68
ZnO sample I9 exciton decay departs
from linearity slightly more than that of 64
ZnO or 66
ZnO samples. This indicates that
for this sample the fast component had a greater relative impact, which implies a
greater influence of surface defects in this sample. This is seen in the Aτf/Aτs ratio in
table A.1. The biexponential nature of the decay in the 68
ZnO is more pronounced in
the I9-1LO data shown in figure A.1(b). Figure A.1(c) shows a comparison of the
TRPL data for the I9 and SX recombinations. The almost single exponential decay
of the I9 exciton is contrasted with the much more pronounced biexponential decay
of the SX, with a very fast initial decay. This is to be expected as the fast component
associated with surface effects is naturally more important for this emission. The
fast component is again fitted with a greater amplitude for the 68
ZnO sample than in
the 66
ZnO and 64
ZnO samples, confirming a greater influence of surface defects in
this sample.
A3
Figure A.1: TRPL decay curves of (a) the I9 recombination, (b) the I9-1LO line, and
(c) the I9 and SX features, in Zn-enriched ZnO nanorods
A4
Table A.1: Fast (τf) and slow (τs) component time constants of the I9 decay lifetime
and their relative amplitudes in Zn-enriched ZnO nanorods.
The TRPL data supports the previous SEM and XRD data in chapter 4 that
the ZnO nanorod samples are of very high quality, although the 68
ZnO sample
appears have a poorer surface quality, as evidenced by the larger fast component
amplitude in TRPL measurements of the biexponential decay, which adversely
affects its TRPL properties. We see no other clear correlation with changes in
isotopic mass in these TRPL measurements.
A.3 Third harmonic UV generation with ZnO nanorods
Non-enriched ZnO nanorods were grown on fused silica substrates using the
NaOH-based chemical bath. The procedure for growing such nanorods was the
same as for Si substrates, as described in section 2.2.3, and resulted in a dense film
of textured, c-axis aligned ZnO nanorods of ~1 μm in thickness. This sample was
then used by our colleagues in Berlin in experiments regarding UV third harmonic
generation (THG).1,2
Figure A.2 shows an image of the CBD-grown nanorods on
these substrates.
Sample 64
ZnO 66
ZnO 68
ZnO
τf (ps) 277.9±8.6 250.3±8.5 223.8±2.1
τs (ps) 698.5±20.3 557.4±19.7 774.2±35.1
Aτf/Aτs 1.21:1.00 1.40:1.00 6.88:1.00
A5
Figure A.2: CBD grown ZnO nanorods on a fused silica substrate. Reproduced
from reference 1.
THG generation has potential application in, to give some examples,
characterisation of ultrashort femtosecond laser pulses and local photodynamic
therapy in the human body.5,6
ZnO nanorods grown by CBD have been used to
study this to the level of single cells.7 This CBD samples were studied in addition to
a VPT samples produced by VLS growth by others, as regards their THG efficiency.
The incident laser had a central wavelength of 810 nm and the THG signal was
observed in both samples at ~ 272 nm. Ultimately, both the CBD sample produced
by the author and the other VPT sample were shown to be efficient sources for UV
THG compared to a bare quartz substrate. This was in contrast to the typical linear
optical characteristics of samples produced by these methods where VPT samples
typically display much higher optical quality than CBD samples (as shown in
chapter 4) and elsewhere.8,9
This study showed the suitability of ZnO nanorods as an
efficient generator of UV emission by up-conversion of red/near infrared laser
pulses. The samples were also shown to be effective in the characterisation of
femtosecond laser pulses.
A6
A.4 Particle and X-ray generation using ZnO nanorods
Figure A.3: ZnO nanorods on Ti foil grown by the (a) NaOH and (b) HMT chemical
baths.
Some ZnO nanorods were also produced during this work for a research visit
undertaken to the Max Born Institute in Berlin in November 2013. For this, ZnO
nanorods were produced using the NaOH chemical bath but using Ti foil as a
substrate. The deposition process was also the same as used with Si substrates.
Same samples of ZnO nanorods were also grown using the HMT method on these
substrates. Coverage on these samples was however less uniform than was typical
for this growth method in our laboratory. The seed layer is most likely the cause of
this as it was more difficult to achieve a smooth and uniform drop-coat of seed
solution on the Ti as it did not wet as easily as the Si. Figure A.3 gives an example
of ZnO nanorods grown by the (a) NaOH and (b) HMT chemical baths on the Ti foil
substrates.
Our colleagues in Berlin used these samples in their study of particle
acceleration with a number of nanostructured targets using high intensity, ultra short,
laser pulses (Ti:Sapphire, 1.3 J, ~30 fs).3 Whether the nanorods, or other
nanostructured morphologies, improve the laser to target coupling, and thus the
particle and x-ray emission efficiency, was investigated.10
The basic principle here
is that lasers at very high relativistic intensities interact with the electrons in the
sample and cause an acceleration of the electron distribution with ions accelerated to
A7
high kinetic energies subsequently. These laser-induced nano-plasmons can
significantly improve the particle acceleration mechanism. Such interaction of high
energy light with matter has attractive opportunities in research and potential
applications. A theoretical model was compared to experimental results and our
CBD samples were among others included in this study. The energy spectral
distributions of protons and carbon ions following an incident laser shot showed that
that the ion numbers were significantly enhanced in our CBD nanorods samples, and
others, in comparison to the plain Ti foil.
A.5 References
1 S.K. Das, F. Güell, C. Gray, P.K. Das, R. Grunwald, and E. McGlynn, Opt. Mater.
Express 4, 701 (2014).
2 S. Kumar Das, F. Güell, C. Gray, D. Byrne, P. Kumar Das, R. Grunwald, G.
Steinmeyer, and E. McGlynn, in Prog. Nonlinear Nano-Optics, edited by S. Sakabe,
C. Lienau, and R. Grunwald (Springer International Publishing, Cham, 2015), pp.
193–206.
3 A. Andreev, K. Platonov, J. Braenzel, A. Lübcke, S. Das, H. Messaoudi, R.
Grunwald, C. Gray, E. McGlynn, and M. Schnürer, Plasma Phys. Control. Fusion
58, 014038 (2016).
4 S. Hong, T. Joo, W. Il Park, Y.H. Jun, and G.-C. Yi, Appl. Phys. Lett. 83, 4157
(2003).
5 Y. Kobayashi, D. Yoshitomi, K. Iwata, H. Takada, and K. Torizuka, Opt. Express
15, 9748 (2007).
6 J. Li, D. Guo, X. Wang, H. Wang, H. Jiang, and B. Chen, Nanoscale Res. Lett. 5,
1063 (2010).
7 S.M. Al-Hilli, M. Willander, A. Ost, and P. Stralfors, J. Appl. Phys. 102, 084304
(2007).
8 D. Byrne, The Growth and Characterisation of Ordered Arrays of Zinc Oxide
Nanostructures and Optical Studies of Defects in Zinc Oxide, Dublin City
A8
University, 2012.
9 D. Byrne, E. McGlynn, J. Cullen, and M.O. Henry, Nanoscale 3, 1675 (2011).
10 M.A. Purvis, V.N. Shlyaptsev, R. Hollinger, C. Bargsten, A. Pukhov, A. Prieto, Y.
Wang, B.M. Luther, L. Yin, S. Wang, and J.J. Rocca, Nat. Photonics 7, 796 (2013).
B1
Appendix B: Efficiency of the SPEX
grating and PMT
As discussed in section 2.4.5, the efficiency curves for the diffraction grating
and PMT used with the SPEX spectrometer, along with their equations, are shown in
figure B.1 and below (reproduced from reference 1).
Figure B.1: SPEX grating and PMT efficiency curves.
Grating efficiency curve:
𝑦 = −59.649 + 0.5479𝑥 − 0.00181𝑥2 + 2.5857 × 10−6𝑥3 − 1.3598 × 10−9𝑥4
PMT efficiency curve:
𝑦 = −1056.763 + 10.6889𝑥 − 0.0424𝑥2 + 8.6543 × 10−5𝑥3 − 9.6602 × 10−8𝑥4
+ 5.621 × 10−11𝑥5 − 1.336 × 10−14𝑥6
B.1 References
1 D. Gorman, Photoluminescence and Excitation Studies of Semiconductors, Dublin
City University, 2001.
C1
Appendix C: XRD Strain
measurements for the Zn-enriched
and VLS O-enriched samples.
As stated in section 5.3.1, between the submission of this thesis for
examination and the viva-voce examination, the XRD strain measurements described
in that section for the VS O-enriched samples were carried out on the Zn-enriched
samples and the VLS O-enriched samples, in order to complete this study on the full
set of samples presented in this work. The peaks positions of the s- and r-planes
with an in-plane component as recorded in these samples and described in section
5.3.1 are contained table C.1. For the Zn-enriched samples, the s-plane peaks vary
by ~0.13° and the r-planes by ~0.1°, although there is no particular trend by mass.
For the VLS O-enriched samples, the s-plane peaks vary by ~0.03° and the r-planes
by ~0.06°, and again there is no particular trend by mass. The actual peak positions
vary slightly from the expected values of 36.256° and 47.541° respectively. This
could be attributed to an instrument calibration which took place in the meantime.
The small shifts in peak positions, and the lack of trend by either average mass or
pure versus mixed samples, indicates there is no discernible strain in our samples.
This is consistent with the result of this experiment for the VS O-enriched samples
discussed in detail in section 5.3.1.
C2
Table C.1: s-plane and r-plane 2θ peaks (with in-plane component) in the Zn-
enriched and VLS O-enriched samples.
Sample
ZnO s-plane
(10-10) peak
2θ (°)
ZnO r-plane
(10-12) peak
2θ (°)
natZnO 36.431 47.739
64ZnO 36.347 47.695
66ZnO 36.453 47.764
68ZnO 36.323 47.679
64/66ZnO 36.427 47.765
66/68ZnO 36.342 47.664
64/68ZnO 36.418 47.751
64/66/68ZnO 36.338 47.695
Zn16
O-VLS 36.386 47.701
Zn16/18
O-VLS 36.365 47.769
Zn18
O-VLS 36.396 47.701